The goal of this study was to utilize a state-of-the-art Speech Emotion Recognition (SER) model to explore the dynamics of basic emotions in semi-structured clinical interviews about depression. Segments of N = 217 interviews from the general population were evaluated using the emotion2vec+ large model and compared with the results of a depressive symptom questionnaire. A direct comparison of depressed and non-depressed subgroups revealed significant differences in the frequency of happy and sad emotions, with participants with higher depression scores exhibiting more sad and less happy emotions. A multiple linear regression model including the seven most predicted emotions plus the duration of the interview as predictors explained 23.7 % of variance in depression scores, with happiness, neutrality, and interview duration emerging as significant predictors. Higher depression scores were associated with lesser happiness and neutrality, as well as a longer interview duration. The study demonstrates the potential of SER models in advancing research methodology by providing a novel, objective tool for exploring emotional dynamics in mental health assessment processes. The model’s capacity for depression screening was tested in a realistic sample from the general population, revealing the potential to supplement future screening systems with an objective emotion measurement.

The importance of cyber-physical-social systems (CPSS) lies in their key role in bridging the gap between the physical, cyber, and social worlds. Integrating topological spaces into CPSS can provide relational structures derived from observations and measurements in the real environment, thereby reducing the risk of accidents and failures when robots operate in social environments. The topological maps of the topological space can be gradually learned by Growing Neural Gas (GNG) from real data, making it suitable for online and real-time applications. However, the training process of GNG encounters difficulties in handling high-density data, leading to significant computational overhead. To address this limitation, this paper proposes a new method to represent the topological space as a multi-layer topological map. This method enables the extraction of information related to different tasks from different topological maps. To enable GNG to effectively learn multiple topologies with different densities, a two-phase approach is proposed: a top-down phase for building a tree structure to accelerate nearest neighbor search, and a bottom-up phase for refining the topology of the upper layer. The experimental results show that it is possible to learn a topology that aligns with the data, overcoming the challenges of conventional methods, and achieving improvements in efficiency and reductions in computational cost, effectively integrating topological space into CPSS.

This paper proposes an Intelligent Evaluation Agent (IEA) with knowledge graph construction based on the Large Language Model (LLM) and Trustworthy AI Dialogue Engine (TAIDE) for personalized Human-Machine Interactive Learning (HMIL). The intelligent agent will deal with multitasks such as learning data preparation and the learner’s data generation, preprocessing, analysis, and evaluation. Multi-modal data is collected from human-machine interactive activities and processed by an IEA to generate structured data stored in human learning repositories. The intelligent agent focuses on various temporal learning periods, such as macro, meso, and micro-level assessments by integrating Human Intelligence (HI) and Machine Intelligence (MI) results, with the MI-based Genetic Algorithm and Neural Network (GANN) learning mechanism employed to optimize the intelligent evaluation model. The learning data evaluation phase aims to identify a model that best fits the group’s learning behavior through HI-based evaluation and to train it further using MI, ensuring that the trained GANN-IEA model closely approximates the HI-based model. An LLM-based knowledge graph agent also supports the evaluation process by helping teachers analyze and visualize students’ learning progress. Experimental results demonstrate that students who study diligently gain knowledge and exhibit increased interest in learning through HMIL. However, the evidence also suggests that some students who excessively rely on Generative AI (GAI) to reproduce learning content without modification become less inclined to engage in diligent study. Additionally, the proposed IEA effectively reduces teachers’ workload in assessing students’ learning status at the end of the semester and supports personalized learning through the designed HMIL model.

Extended Reality, artificial intelligence, and big data technologies offer new opportunities for advancing rehabilitation diagnosis and training. This study presents a method for extracting behavioral features from a visual search task conducted in an immersive virtual reality environment. To identify motion patterns specific to individual patients, we employ a topological mapping approach based on Growing Neural Gas (GNG), which adapts its structure dynamically using node activation and error-based edge management. While GNG effectively captures spatial characteristics, it lacks the ability to model temporal relationships and is sensitive to hyperparameter settings. To address these limitations, we introduce a spatiotemporal topological clustering method, along with a hierarchical framework that enables segmentation at multiple levels of granularity. Furthermore, to evaluate feasibility, we conducted a visual search task with three patients, including one with USN, and performed a comparative analysis of their extracted motion features.

Event cameras offer superior performance in terms of temporal resolution, dynamic range, latency, and power efficiency compared to conventional cameras. However, due to their unique operational principles, which involve measuring changes in luminance values, tasks such as human pose estimation become more challenging. In this study, we propose a real-time feature extraction method from event data for human pose estimation using event cameras. Specifically, after adjusting resolutions with Multi-resolution Maps, we employ Growing Neural Gas for topological clustering. Subsequently, spectral clustering is performed based on the resulting graph structure, enabling the segmentation of human body parts with a focus on the graph structure.

|Unilateral Spatial Neglect (USN) is characterized by an attention deficit to one side of space, where individuals struggle to perceive stimuli on that side without a lack of sensation. Traditional paper-pencil tasks like line cancellation and copying tests are commonly used to assess USN, but they have limitations in evaluating neglect areas confined to a two-dimensional plane. Immersive VR systems and multimodal sensing systems offer a more sensitive approach for diagnosis and training. In related works, AR/VR systems and eye-tracking devices are utilized for measuring, evaluating, and creating assessment tasks for USN. However, these approaches can only analyze relationships between perception and movements in specific environments. In this study, we propose a method for cognitive modeling based on perceiving-acting cycle in USN, utilizing computational intelligence techniques to establish a structured coupling framework, aiming to contribute to a novel and effective approach for understanding and addressing USN.

To support humans in various environments such as homes, nursing homes, and hospitals, robots need a spatial perception system that enables flexible and rapid recognition of unlearned objects. However, there is a problem that methods that require prior learning cannot fully deal with unlearned objects and that the input of RGB images is easily affected by the lighting conditions of the environment. In this study, a topological map is generated in real-time using Growing Neural Gas, a type of unsupervised self-growing neural network, with 3D point cloud data as the only input. However, in unknown environments, there is a problem that the size of clusters becomes unstable and cannot be detected stably if the region growing method, which checks only labels assigned based on normal vectors, is used to detect clusters such as unknown objects in the topological map. Therefore, we propose a methodology to stably detect clusters in a topological map by introducing the concept of age and a reliability based on fuzzy sets for each node and rejecting unstable nodes with low reliability even if the nodes to which they are connected have the same label. Also, the effectiveness of the proposed methodology is demonstrated by experiments through 3D dynamics simulations.

This paper presents a method for human motion segmentation aimed at motion analysis in healthcare and rehabilitation. Motion segmentation involves extracting small movements, known as motion primitives, from a sequence of behavioral patterns. While previous works have utilized unsupervised clustering methods as effective approaches for motion segmentation, many of these methods require prior knowledge to enhance performance. To overcome these challenges, we propose a hierarchical topological clustering method capable of representing spatiotemporal features using GNG and the Pulse Neuron Model. Additionally, we present experiments and discussions to validate the applicability of the proposed method for motion analysis in exercise.

Abstract In this journal, Cheng has proposed a backpropagation (BP) procedure called BPFCC for deep fully connected cascaded (FCC) neural network learning in comparison with a neuron-by-neuron (NBN) algorithm of Wilamowski and Yu. Both BPFCC and NBN are designed to implement the Levenberg-Marquardt method, which requires an efficient evaluation of the Gauss-Newton (approximate Hessian) matrix $$\nabla \textbf{r}^\textsf{T} \nabla \textbf{r}$$ ∇ r T ∇ r , the cross product of the Jacobian matrix $$\nabla \textbf{r}$$ ∇ r of the residual vector $$\textbf{r}$$ r in nonlinear least squares sense. Here, the dominant cost is to form $$\nabla \textbf{r}^\textsf{T} \nabla \textbf{r}$$ ∇ r T ∇ r by rank updates on each data pattern. Notably, NBN is better than BPFCC for the multiple $$q~\!(>\!1)$$ q ( > 1 ) -output FCC-learning when q rows (per pattern) of the Jacobian matrix $$\nabla \textbf{r}$$ ∇ r are evaluated; however, the dominant cost (for rank updates) is common to both BPFCC and NBN. The purpose of this paper is to present a new more efficient stage-wise BP procedure (for q-output FCC-learning) that reduces the dominant cost with no rows of $$\nabla \textbf{r}$$ ∇ r explicitly evaluated, just as standard BP evaluates the gradient vector $$\nabla \textbf{r}^\textsf{T} \textbf{r}$$ ∇ r T r with no explicit evaluation of any rows of the Jacobian matrix $$\nabla \textbf{r}$$ ∇ r .

In recent years, exercise recognition has become increasingly popular for exercise monitoring and rehabilitation for older adults. However, to identify non-learned exercises, another dataset needs to be collected for differentiation purposes. This study aims to provide an add-on technique to the encoder model to distinguish between learned and non-learned exercises without training with non-learned exercise data. First, we form a list of activation nodes based on the output of the encoder by using the k-means algorithm. Afterward, these nodes are used to compute activation scores from encoded features. These activation scores are used to differentiate non-learned exercises by a threshold value. After differentiation, the activation scores are then passed to a multi-layer perceptron (MLP) for exercise classification. Meanwhile, we proposed a unique method to compute a distinguishing score to find the optimal$k-\mathbf{nodes}$. We demonstrate the proposed method using the MM-Fit dataset, showing that it can identify fitness exercises and distinguish non-learned exercises without much performance loss. Lastly, we collected a dataset about Chair-Fitness activity to validate the proposed method's effectiveness further and welcome other researchers to utilize the dataset.

To realize a super-smart society (Society 5.0) where humans and robots coexist, there is a need for a perceptual system that can recognize unknown objects in various unknown environments quickly and flexibly. In unknown environments, the characteristics of objects cannot be known in advance, so prior learning-based recognition methods such as deep reinforcement learning cannot fully cover the problem. There have been many studies on environment recognition (clustering, etc.) using a combination of RGB images and distance images, but the recognition performance is unstable because it strongly depends on the lighting conditions of the environment. Therefore, in this study, we construct a 3D topological map of the environment in real-time using Growing Neural Gas (GNG), which can learn 3D topological structures even for unlearned objects, using only 3D point cloud data as input. In the real world, due to the characteristics of RGB-D cameras, sample density decreases for more far-away objects and only sparse depth information can be obtained, so conventional GNG cannot generate high-density topological structures of unknown objects. Therefore, if the object category labels of the winner nodes (nearest nodes) for the input vector (3D point cloud) match the unknown object and are within a predefined tolerance area, then it is judged to be useful input information for learning the topological structure of the unknown object, and the topological structure of the unknown object is learned. We propose Add-if-Silent rule-based GNG (AiS-GNG) which can generate high-density topological structures for far-away objects by directly adding input data as a reference vector. We verify the effectiveness of the proposed method through experiments using a 3D dynamics simulator.

We aim to develop a simple motion capture system for home environments. As users need to install their own cameras, a calibration-free system is required. Therefore, we propose a 3D pose estimation method based on 3D joint angles of humans using multiple smart devices with a hybrid optimization method that combines Particle Swarm Optimization and steepest descent method. We also estimate the relative angles between humans and cameras to facilitate camera calibration. In this paper, we discuss the impact of the combination of global and local search capabilities of the optimization method on the system's performance. Specifically, we propose an optimization method that gradually changes the number of iterations of Particle Swarm Optimization and Steepest Descent Method and compare it with a simple sequential combination.

Theaim of this study is to model the cognitive processes based on a perceiving-acting cycle in patients with unilateral spatial neglect (USN). USN is the inability to perceive features of the environment, body, or objects on one side. To extract the cognitive characteristics of USN patients in a multifaceted manner, we constructed a multimodal sensing system using immersive VR. In this paper, we present a system that predicts movement of a subject while performing a search task using the measurement results and an LSTM neural network.

In order to realize a super-smart society (Society 5.0) where humans and robots coexist, there is a need for a perceptual system that can recognize the environment quickly and flexibly in an environment that changes from moment to moment. In an unknown environment, the characteristics of objects cannot be known in advance, and thus prior learning-based recognition methods such as deep reinforcement learning may not be able to cope with this situation. In this study, we construct a 3D topological map of the environment in real-time using Growing Neural Gas (GNG), which can learn 3D topological structures even for unlearned objects. However conventional GNG have the problem that they cannot generate nodes with high-density for distant objects and cannot identify whether an unknown object is static or dynamic. Therefore, by directly adding useful input data as a new node (reference vector) based on the object category labels of the winner nodes (nearest nodes) to the input vector (3D point cloud), it is possible to generate high-density topological structures even for distant objects. We proposed the Add-if-Silent rule-based GNG with Amount of Movement (AiS-GNG-AM), which can identify between static and dynamic objects based on the past amount of movement of a node. The effectiveness of the proposed method is verified through experiments using a 3D dynamics simulator.

With the emergence of an ultra-smart society, it is desirable to have a mobility support robot that can act flexibly like a human without prior knowledge or trial-and-error learning. One solution is to appropriately judge the space to be paid attention to and make decisions using only time-series observation information of the surrounding area in public spaces where many people come and go freely. Our goal is to use a spatial-attention-based sensory network for a fuzzy controller to adapt to dynamic environments and safely perform complex navigation tasks using only time-series observation information, without requiring prior knowledge such as the type, speed, and direction of moving obstacles, and more importantly, without trial-and-error learning or control system redesign. The computer simulation results show that the proposed spatial attention-based sensory network and situation-based behavior coordination allow the mobility support robot to adapt online to a complex dynamic environment with multiple moving obstacles. In addition, the proposed method can perform the navigation task more safely and efficiently than the conventional method.

Hand rehabilitation in post-stroke patients with visual impairment is currently not supported by the availability of a cyber-physical-social system (CPSS) that can monitor physical development during daily activities. This paper discusses how to extract hand activity information on objects based on visual attention in the task-specific reach-to-grasp cycle. We used perception-based egocentric vision to observe hand-object interactions (HOI) in grasping tasks. Our approach combines object detection with hand skeletal model estimation and visual attention to validate HOI detection. We choose a multilayer Gated Recurrent Unit (GRU) based on Recurrent Neural Networks (RNN) architecture to classify the four main activities when the hand interacts with an object (wonder-reach-grasp-release). We evaluated the algorithm quantitatively on the new dataset we introduced for cup grasping activity. This method can validate the HOI detection with 97.0%precision with less training time for small data. Further research will use these results to increase self-efficacy for independent hand-eye coordination rehabilitation support in community-centric systems. The code and dataset are available at https://github.com/anom-tmu/hoi-attention/.

The general success criterion for an artificial intelligence system is its ability to mimic human brain learning. Throughout a lifetime, the human brain is capable of continual learning. The acquired information is kept, augmented, finetuned, and utilized to complete new tasks in the future. At the moment, machine learning models perform well when given precisely structured, balanced, and homogenized data. However, when several jobs with incremental data are provided, the performance of the majority of these models suffers. Inspired by the Complementary Learning Systems (CLS) theory in neuroscience, episodic-semantic memory-based frameworks have received much attention and research. On the other hand, conventional methods are needed to perform data batch normalization and are sensitive to vigilance hyperparameters across different datasets. This paper proposes a Robust Growing Memory Network (RGMN) that continuously learns incoming data without normalization and is unlikely to be affected by the vigilance hyperparameter. The RGMN is a self-organizing topological network that models human episodic memory, and its network size can grow and shrink in response to data. The long-term memory buffer retains the largest and smallest data values that will use for learning. To evaluate the performance of the proposed method, we conducted comparative experiments on real-world datasets, and results showed that the proposed method outperforms existing memory-based baseline frameworks in terms of accuracy.

In recent years, advances in information and communication technology have led to the research and development of cyber-physical systems and digital twin that simulate various events in real space in cyber space. In the field of human flow simulation, it is possible to predict and analyze the flow of people in various spaces, from indoor to outdoor, and use this information to create spaces that promote human activities, such as searching for optimal layouts and alleviating congestion by distributing traffic lines. In this paper, we propose to use multi-scopic simulation to simulate human flow. Next, using the human flow data measured by the simulation, we extract and analyze the features by topological mapping. Finally, we discuss the effectiveness of the proposed method through some simulation results.

This papers describes a learning algorithm for growing neural gas to construct a topology-preserving map from a 3D point cloud whose topology can change dynamically. Growing Neural Gas with Utility Factor (GNG-U) has been presented as a method for learning the topology of a 3D space environment and applying it to non-stationary or dynamic data distribution. However, when a node is added to an existing network after several errors with sampling data have accumulated, it is difficult for a standard GNG-U to considerably boost learning speed. As a result, we propose a revolutionary growth strategy that dramatically accelerates learning and convergence. This method immediately adds a sample of data as a new node to an existing network based on the likelihood of node addition estimated by the distance to the third closest node and the first and second closest nodes at maximum. Experiment findings show that the proposed algorithm’s network can quickly adapt to represent the topology of non-stationary input distributions.

To reduce the number of traffic fatalities, research and development of safety systems is currently under way. However, drivers are still required to judge traffic situations and operate vehicles safely. Therefore, the aim of this study is to clarify the effect of pedestrian information provided by a human–machine interface (HMI) on driving characteristics under different traffic conditions when vehicles turn right at intersections and drive straight where a pedestrian is crossing. The system is designed to present information on pedestrians to the drivers through the HMI. The experimental results reveal that the presentation of pedestrian information by the HMI increases the time to collision (TTC) and decreases closest distance to the pedestrian at intersections and on straight roads. Thus, it is effective for drivers to ensure safety. In contrast, results reveal that the presentation by the HMI has no effect on increasing the TTC or decreasing the closest distance to the pedestrian in the two types of driving situations; one case where the ego vehicle makes a right turn at the intersection just after the oncoming vehicle turns left, and another case where the ego vehicle goes straight with good visibility. Therefore, for a more effective HMI, a suitable design or timing of pedestrian information should be considered, depending on various traffic environments.

This article examines the importance of integrating locomotion and cognitive information for achieving dynamic locomotion from a viewpoint combining biology and ecological psychology. We present a mammalian neuromusculoskeletal model from external sensory information processing to muscle activation, which includes: 1) a visual-attention control mechanism for controlling attention to external inputs; 2) object recognition representing the primary motor cortex; 3) a motor control model that determines motor commands traveling down the corticospinal and reticulospinal tracts; 4) a central pattern generation model representing pattern generation in the spinal cord; and 5) a muscle reflex model representing the muscle model and its reflex mechanism. The proposed model is able to generate the locomotion of a quadruped robot in flat and natural terrain. The experiment also shows the importance of a postural reflex mechanism when experiencing a sudden obstacle. We show the reflex mechanism when a sudden obstacle is separately detected from both external (retina) and internal (touching afferent) sensory information. We present the biological rationale for supporting the proposed model. Finally, we discuss future contributions, trends, and the importance of the proposed research.

In the case of vehicles with low speeds at the time of pedestrian fatality, the percentage of pedestrian collisions was the highest for right turns, yet the mechanism of these traffic accidents has not been clarified. In this study, we investigate the behavioral characteristics of drivers when a vehicle makes a right turn in five situations using a driving simulator. We conducted an experiment using a driver assistance system that alerted drivers when the system detected pedestrians at the intersection. A human–machine interface (HMI) was first displayed when the subject vehicle (ego vehicle) stopped in front of the intersection due to a red light. The display was then turned off when the traffic light changed to green, and the ego vehicle started moving. It was displayed again when the ego vehicle entered the intersection. We found that HMI display was effective in increasing the percentage of driver’s gazing time at pedestrians and in ensuring safety by the vehicle’s stopping to move forward. Furthermore, we found that HMI’s effectiveness was the most significant in the situation when three preceding vehicles made a right turn.

Recently, as the rapid progress of information and communication technology, robot technology, and artificial intelligence, we have become to build a higher level of safe, comfortable, and smart society coexisting with advanced technology. Methamphetamine addiction has become a major human social problem in the world. Traditional approaches of detecting methamphetamine through hair, skin, urine, and blood fluid are financially inefficient, time-consuming, and in some cases too complicated. Providing reliable and trustworthy detection with the highest precision and accuracy is of a challenging task. This paper proposes ensemble learning using a soft voting approach to improve the accuracy of detection. First, we trained five individual classifiers, namely adaptive neuro-fuzzy inference system (ANFIS), random forest, multilayer perceptron (MLP), k-nearest neighbor (k-NN), and support vector machine (SVM) on the same urine dataset. We then created new ensemble learning using the soft voting approach by averaging the probability of individual ANFIS, random forest, MLP, k-NN, and SVM. Firefly algorithm for weight optimization is used to strengthen individual classifiers to form an ensemble and increase the prediction accuracy. Our proposed ensemble produces an accuracy value of 100% compared to the individual classifiers mentioned above.

In this paper, we propose an AI-FML agent for robotic game of Go and AIoT real-world co-learning applications. The fuzzy machine learning mechanisms are adopted in the proposed model, including fuzzy markup language (FML)-based genetic learning (GFML), eXtreme Gradient Boost (XGBoost), and a seven-layered deep fuzzy neural network (DFNN) with backpropagation learning, to predict the win rate of the game of Go as Black or White. This paper uses Google AlphaGo Master sixty games as the dataset to evaluate the performance of the fuzzy machine learning, and the desired output dataset were predicted by Facebook AI Research (FAIR) ELF Open Go AI bot. In addition, we use IEEE 1855 standard for FML to describe the knowledge base and rule base of the Open Go Darkforest (OGD) prediction platform in order to infer the win rate of the game. Next, the proposed AI-FML agent publishes the inferred result to communicate with the robot Kebbi Air based on MQTT protocol to achieve the goal of human and smart machine co-learning. From Sept. 2019 to Jan. 2020, we introduced the AI-FML agent into the teaching and learning fields in Taiwan. The experimental results show the robots and students can co-learn AI tools and FML applications effectively. In addition, XGBoost outperforms the other machine learning methods but DFNN has the most obvious progress after learning. In the future, we hope to deploy the AI-FML agent to more available robot and human co-learning platforms through the established AI-FML International Academy in the world.

Recently, various types of communication robots have been developed all over the world to make up for a lack of human labor. The purpose of communication robots is to provide services such as information support including facility guide and recommendation at airports and shopping centers, nursing homes for elderly people. In general, the effective integration of various modules is important to aim for rapid practical application in an aging society. However, it takes much cost to introduce communication robots because of software customization and contents design for information service in addition to the maintenance cost of software and hardware. Furthermore, we have to pay attention to safe physical interaction and considerable communication of communication robots with people. In this paper, we develop a robot partner for information support and propose a new method to recommend various information flexibly according to human intention.

In recent years, the number of elderlies living alone has been increasing year by year due to the declining birthrate and aging problem, which has become a major problem especially in developed countries. There is a great demand for health support systems and monitoring systems using sensors to support the elderly in such situations. In this paper, we propose a robot partner system that can communicate based on sensor data in such a situation.

In recent years, the number of elderly people over the age of 65 who live alone has increased year by year. In particular, the frequency of conversation for the elderly tends to decrease to less than once in two weeks, and the frequency of going out from home tends also to decrease at a rate of less than once in 2-3 days. As a result, it has led to the social isolation of the elderly, and has caused social problems such as progression of dementia and lonely death [1]. For that reason, it is necessary to provide support to increase the frequency of conversation among isolated elderly people. Accordingly, researches on human robot interaction are attracting attention. The interaction with robot partners could create a good stimulating environment so that the elderly could communicate with their family and others. Therefore, we focus on the interaction between humans and robots, and develop a robot system that focuses on increasing the frequency of conversations caused by touching. In this paper, we explain the robot partner system that used touch sensory information. Then, we show experimental results of the effectiveness of robot partner system, and discuss the applicability of the proposed system.

Falls are a major health issue, particularly among the elderly. Increasing fall events require high service quality and dedicated medical treatment which is an economic burden. In the lack of appropriate care and support, serious injuries caused by fall will cost lives. Therefore, tracking systems with fall detection capabilities are required. Static-view sensors with machine learning techniques for human fall detection have been widely studied and achieved significant results. However, these systems unable to monitor a person if he or she is out of viewing angle which greatly impedes its performance. Mobile robots are an alternative for keeping the person in sight. However, existing mobile robots are unable to operate for a long time due to battery issues and movement constraints in complex environments. In this paper, we proposed a lightweight deep learning vision-based model for human fall detection with an assistive robot to provide assistance when a fall happens. The proposed detection system requires less computational power which can be implemented in a low-cost 2D camera and GPU board for real-time monitoring. The assistive robot equipped with various sensors that can perform SLAM, obstacle avoidance and navigation autonomously. Our proposed system integrates these two sub-systems to compensate for the weakness of each other to constitute a system that robust, adaptable, and high performance. The proposed method has been validated through a series of experiments.

Human and animal locomotion are controlled by complex neural circuits, which can also serve as inspiration for designing locomotion controllers for dynamic locomotion in legged robots. We develop a locomotion controller model including a central pattern generator (CPGs) and a muscle reflex based on the forelimb and hindlimb structures of a cat. In this paper, we focus on modeling the muscle reflex and its optimization. This muscle reflex model regulates ground force afferents in each limb. There are two phases in each step performed by this model, the swing and stance phases. The muscle during swing phase is activated by a pattern formation signal from the CPG. During stance phase, the muscle is automatically controlled by the moving speed. We utilize a multi-objective evolutionary algorithm to optimize parameters of the model. We use the proposed model to control a cat-like robot in simulations using Open Dynamics Engine. Results show that the simulated robot is able to move at different speeds by modulating simple stimulation signals to the CPG without needing to modify muscle and reflex parameters.

In the field of quadruped locomotion, dynamic locomotion behavior, and rich integration with sensory feedback represents a significant development. In this paper, we present an efficient neural model, which includes CPG and its sensorimotor coordination, and demonstrate its implementation in a quadruped robot to show how efficient integration of motor and sensory feedback can generate dynamic behavior and how sensorimotor coordination reconstructs the sensory network for leg malfunction compensation. Additionally, we delineate a network optimization strategy and suggest sensorimotor coordination as a strategy for controlling speed and regulating internal and external adaptation. The rhythm generation representing the leg injury was inactive, stimulating the sensorimotor system to reconstruct the network between CPG and feet force afferent without any commanding parameter. The performances of the simulated and real, cat-like robot on both flat and rough terrains and the leg malfunction tests demonstrated the effectiveness of the proposed model, indicating that a smooth gait-pattern transition could be generated during sudden leg malfunction.

In this paper, we propose an AI-FML agent with a patch learning (PL) mechanism for the robotic game of Go applications. The proposed AI-FML agent contains three kinds of intelligence, including a perception intelligence, a cognition intelligence, and computational intelligence, for the robotic application. Additionally, we embed the PL mechanism into the AI-FML agent. The method for running PL involves three steps. It first trains an initial global model, then trains a patch model for each identified patch, and finally updates the global model using the training data that do not fall into any patch. This paper adopts the Google DeepMind Master 60 games to be the training data and testing data set. The experimental results show the AI-FML agent with the patch learning mechanism can improve the performance of regression for the robotic game of Go applications.

Recently, various types of robot partners aiming at social implementation have been developed and researched, but they have not yet been popularized. In the case of communication robots, it has been reported that it is easier and more comfortable for humans to communicate with multiple robots. However, multirobots require large installation space; the role or character of each robot is not clear, and it is difficult for users to design utterance and action contents. Therefore, we propose a new design methodology of the body-sharing multi-robot system and develop a kangaroo-based robot that contains two physically interconnected objects as one of the models of multiple communication robots. Next, we show the evaluation of the proposed robot in terms of impression and discuss the possibility of social implementation of the developed body sharing multirobot system.

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record.This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article.Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

The currently observed developments in Artificial Intelligence (AI) and its influence on different types of industries mean that human-robot cooperation is of special importance. Various types of robots have been applied to the so-called field of Edutainment, i.e., the field that combines education with entertainment. This paper introduces a novel fuzzy-based system for a human-robot cooperative Edutainment. This co-learning system includes a brain-computer interface (BCI) ontology model and a Fuzzy Markup Language (FML)-based Reinforcement Learning Agent (FRL-Agent). The proposed FRL-Agent is composed of (1) a human learning agent, (2) a robotic teaching agent, (3) a Bayesian estimation agent, (4) a robotic BCI agent, (5) a fuzzy machine learning agent, and (6) a fuzzy BCI ontology. In order to verify the effectiveness of the proposed system, the FRL-Agent is used as a robot teacher in a number of elementary schools, junior high schools, and at a university to allow robot teachers and students to learn together in the classroom. The participated students use handheld devices to indirectly or directly interact with the robot teachers to learn English. Additionally, a number of university students wear a commercial EEG device with eight electrode channels to learn English and listen to music. In the experiments, the robotic BCI agent analyzes the collected signals from the EEG device and transforms them into five physiological indices when the students are learning or listening. The Bayesian estimation agent and fuzzy machine learning agent optimize the parameters of the FRL agent and store them in the fuzzy BCI ontology. The experimental results show that the robot teachers motivate students to learn and stimulate their progress. The fuzzy machine learning agent is able to predict the five physiological indices based on the eight-channel EEG data and the trained model. In addition, we also train the model to predict the other students’ feelings based on the analyzed physiological indices and labeled feelings. The FRL agent is able to provide personalized learning content based on the developed human and robot cooperative edutainment approaches. To our knowledge, the FRL agent has not applied to the teaching fields such as elementary schools before and it opens up a promising new line of research in human and robot co-learning. In the future, we hope the FRL agent will solve such an existing problem in the classroom that the high-performing students feel the learning contents are too simple to motivate their learning or the low-performing students are unable to keep up with the learning progress to choose to give up learning.

This Special Section collects the latest developments in video system design, data compression, target detection, object localization, behavior analysis, motion detection, and real-time implementation of industrial video systems and intelligent analytics to bring the latest ideas and solutions of the research community on practical video systems to our audience. The Special Section focuses on several topics that are recently concerned in the community, including, multicamera network, real-time hardware implementation, networked data analytics, bandwidth limited compression, motion pattern analysis, action understanding, three-dimensional (3-D) reconstruction, contextual recognition, object detection and tracking, intelligent robot vision, security surveillance, intelligent transportation, and other industrial applications. The Special Section presents 14 articles on intelligent video systems and analytics. These articles are briefly summarized here.

This paper applies a genetic algorithm and fuzzy markup language to construct a human and smart machine cooperative learning system on game of Go. The genetic fuzzy markup language (GFML)-based Robot Agent can work on various kinds of robots, including Palro, Pepper, and TMU's robots. We use the parameters of FAIR open source Darkforest and OpenGo AI bots to construct the knowledge base of Open Go Darkforest (OGD) cloud platform for student learning on the Internet. In addition, we adopt the data from AlphaGo Master's sixty online games as the training data to construct the knowledge base and rule base of the co-learning system. First, the Darkforest predicts the win rate based on various simulation numbers and matching rates for each game on the OGD platform, then the win rate of OpenGo is as the final desired output. The experimental results show that the proposed approach can improve knowledge base and rule base of the prediction ability based on Darkforest and OpenGo AI bot with various simulation numbers.

In recent years, demands for service robots are getting bigger as the decrease in population of the labor declines. In this paper, we will develop and research about system of the guide robot with moving mechanism among service robots. The guide robot in this research aims to lead a user receiving guidance and autonomously move to the target position.

Biological information inspires the advancement of a navigational mechanism for autonomous robots to help people explore and map real-world environments. However, the robot's ability to constantly acquire environmental information in real-world, dynamic environments has remained a challenge for many years. In this paper, we propose a self-organizing adaptive recurrent incremental network that models human episodic memory to learn spatiotemporal representations from novel sensory data. The proposed method termed as SOARIN consists of two main learning process that is active learning and episodic memory playback. For active learning (robot exploration), SOARIN quickly learns and adapts incoming novel sensory data as episodic neurons via competitive Hebbian Learning. Episodic neurons are connecting with each other and gradually forms a spatial map that can be used for robot localization. Episodic memory playback is triggered whenever the robot is in an inactive mode (charging or hibernating). During playback, SOARIN gradually integrates knowledge and experience into more consolidate spatial map structures that can overcome the catastrophic forgetting. The proposed method is analyzed and evaluated in term of map learning and localization through a series of real robot experiments in real-world indoor environments.

At the beginning of the research on artificial intelligence, intelligence was thought to be almost synonymous with thought, but as the research of psychology and robot progresses, it began to be considered that the essence of intelligence is to have physicality. In research on cognitive developmental robotics, various methodologies related to cognitive developmental learning by communication-based on physicality are proposed, and even in human cognitive development, cognitive ability, communication skill, and physical exercise ability interrelate with each other is important. In this paper, for instructed learning between humans and humanoid robots, we focus on the analysis of nonverbal information and construct a cognitive development the system with the objective of recognizing the target motion, extracting action segments, and acquiring actions. We propose a method to recognize the motion of human beings in real time by using Spiking Neural Network, cluster direction vectors using Multi Layer Growing Neural Gas, and obtain motion. In addition, in order to confirm the effectiveness of the proposed method, actions are acquired using data actually measured, and the acquired action is confirmed.

This paper presents a method with dynamic density topological structure generation for low-cost real-time vertical ladder detection from 3D point cloud data. Dynamic Density Growing Neural Gas (DD-GNG) is proposed to generate a dynamic density of the topological structure. The density of the structure and the number of nodes will be increased in the targeted object area. Feature extraction model is required to classify suspected objects for being processed in the next time process. After that, rungs of the vertical ladder is processed using an inlier-outlier method. Thus, the ladder detection model represents the ladder with a set of nodes and edges. Next, affordance detection is processed for detecting the feasible grasped location. To validate the effectiveness of the proposed method, a series of experiments are conducted on a 4-legged robot with a non-GPU board for real-time vertical ladder detection and climbing to validate the effectiveness of the proposed method. Results show that our proposed method able to detect and track the ladder structure in real-time with a much lower computational cost. The affordance of the ladder provides safety information for robot grasping.

This paper presents the novel capabilities of a quadruped robot by performing horizontal-vertical-horizontal movement transition through vertical ladder without handrailing supporter. To overcome the proposed problem, we propose a multi-behavior generation model using independent stepping and pose control in the quadruped robot. The model is able to generate appropriate behavior depending on external (3D point clouds) and internal sensors (ground touch sensor, and Inertial Measurement Unit). Posture condition, safe movement area, possible touchpoint, grasping possibility, and target movement are the information that is analyzed from the sensors. There are four options developed in behavior generation, which are, Approaching, Body Placing, Stepping, and Grasping behavior. In order to prove the effectiveness of the proposed algorithm, the model was implemented on the computer simulation and the real application. Before being applied in the real robot, the proposed model is optimized in the computer simulation. Then, the optimized parameter is used for applying in the real robot. As a result, the robot succeeded to move through the ladder without handrail from lower stair to upper stair. From the analysis, the body placing behavior is the most important strategy in the proposed case.

This paper proposes layered neural based locomotion with a hierarchical learning process. Central pattern generation (CPG) in a higher layer generates an analog signal to the lower layer which is Motor Neurons Pools. In a lower layer, motor neuron generates the angular velocity of the joint. Central pattern generation is built based on neural oscillator model in the spinal cord. It responds to the pattern of locomotion model. Then, the inner state of a motor neuron is developed based on muscle activity in the human musculoskeletal model. Furthermore, the motor neurons pools (MNs) are integrated with sensory neurons (SNs) that send the internal feedback of the robot. There are two steps of the learning process: 1) optimizing the structure of CPG for generating appropriate behavior without considering the feedback information 2) optimizing the integration between MNs and SNs for generating adaptive behavior toward internal disturbance. The proposed model has been implemented in a simulated humanoid robot with carrying different weight of payloads. The robot behavior changes for stabilizing the robot posture.

By using a combination of vision and depth map sensors, this paper aims at detecting the real-time affordability of gripping pose for hand-over object behavior in robot-human interaction. The affordance detection will serve a set of seven-dimension gripping information (3D location, 3D Rotation, and gripping size). Moreover, the novelty is that the goal has to consider gripping behavior of the receiver. Therefore, the result of the proposed method discusses the gripping location of the robot and the estimation of the receivers gripping location. Technically, desired objects detection is computed using a computer vision algorithm. After that, from the depth information, the topological map will be generated using the proposed dynamic density growing neural gas. The density topological structure will be focusing on the desired object. With the topological map information and robot gripper embodiment, the possible gripping position is computed based on the inlier-outlier method. Ranking information in every detected gripping is also considered for classifying from the best and the worst gripping position. Experimental results showed that the proposed work capable of detecting the affordance with a gripping recommendation in real-time with a low computational cost.

Studies have shown that people relate better with other people who have similar personality characteristics as themselves. This is helpful in peer support scenarios where people should be receptive to receiving support and advice from others. In this paper we propose a chatbot personality model and an algorithm that enables the chatbot to adapt its personality in real-time as it interacts in conversation with the user. Our model is based on the Big Five personality model and we focus on two key personality traits, Extroversion and Agreeableness. The personality adaption algorithm uses an interactive genetic algorithm. We have exposed the chatbot to a controlled set of interactions and a user and the results show that the algorithms are capable of adapting personality traits to match the identified traits of the user. In the next phase of our experimentation we will expose the chatbot to healthcare professionals.

An intelligent robot agent based on domain ontology, machine learning mechanism, and Fuzzy Markup Language (FML) for students and robot co-learning is presented in this paper. The machine-human co-learning model is established to help various students learn the mathematical concepts based on their learning ability and performance. Meanwhile, the robot acts as a teacher's assistant to co-learn with children in the class. The FML-based knowledge base and rule base are embedded in the robot so that the teachers can get feedback from the robot on whether students make progress or not. Next, we infer students' learning performance based on learning content's difficulty and students' ability, concentration level, as well as teamwork spirit in the class. Moreover, we combine the optimization techniques such as genetic algorithm (GA) and particle swarm optimization (PSO) with FML, called GFML and PFML, respectively, to learn the constructed knowledge base and rule base. Experimental results show that learning with the robot is helpful for disadvantaged and below-basic children. Moreover, the accuracy of the intelligent FML-based agent for student learning is increased after machine learning mechanism.

The use of electronic textbooks (e-book) has been heavily studied over the years due to their flexibility, accessibility, interactivity and extensibility. Yet current shortcomings of e-book, which is often just a digitized version of the original book, do not encourage adoption. Consequently, this leads to a rethinking of e-book that should incorporate current technologies to augment its capabilities, where inclusion of information search and organization tools have shown to be favorable. This paper is on a preliminary work to add intelligence into such tools in terms of information retrieval. Construction of knowledge graph for e-book material with little overhead is first introduced. Information retrieval through typed similarity query is then performed via random walk. Case study demonstrates the applicability of the e-book platform, with promising application and advancement in the area of electronic textbooks.

In this paper, we present a novel biologically inspired evolving neural oscillator for quadruped robot locomotion to minimize constraints during the locomotion process. The proposed sensory-motor coordination model is formed by the interconnection between motor and sensory neurons. The model utilizes Bacterial Programming to reconstruct the number of joints and neurons in each joint based on environmental conditions. Bacterial Programming is inspired by the evolutionary process of bacteria that includes bacterial mutation and gene transfer process. In this system, either the number of joints, the number of neurons, or the interconnection structure are changing dynamically depending on the sensory information from sensors equipped on the robot. The proposed model is simulated in computer for realizing the optimization process and the optimized structure is then applied to a real quadruped robot for locomotion process. The optimizing process is based on tree structure optimization to simplify the sensory-motor interconnection structure. The proposed model was validated by series of real robot experiments in different environmental conditions.

Human actions can be realized by observing the trajectories of skeleton joints. In this paper, we propose an unsupervised episodic memory learning model for skeleton based action learning and recognition. The proposed model, Multi-channel Episodic Memory Adaptive Resonance Theory (McEM-ART), consists of three layers: short term memory, working memory and Episodic memory. The short term memory layer is formed by multiple ART networks to obtain sensory data and cluster them into neurons in working memory layer. Instead of obtaining the whole skeleton as an input, we divide the human skeleton into three parts, upper part body, main body and lower part body. Each of them is then feed to McEM-ART short term memory layer for learning. Episodic memory layer extracts novel events and encodes spatio-temporal connection between them as episodes by generating cognitive neurons incrementally for action recognition. Comparing with previous works, McEM-ART further integrates a novel memory anticipation functions for encoding crucial events and episodes and recalling them using partial and inexact cues. Experimental results demonstrate that McEM-ART is capable of clustering human skeleton data into event neurons, encoding sequence of activation events as episode neurons for action recalling and recognition.

In recent years, various kinds of communication robots are active in our lives. However, most of them consider only the existence of speaker and listener. In this paper, we aim to realize a more lively and lifelike communication by developing the relationship between speaker, listener and objects that exist in the environment. We propose a cognitive environment system for a robot partner to determine objects that referred by humans. The proposed system measures the degree of relevance of objects existing in the environment from two perspectives such as the verbal information based on relevance theory and non-verbal information based on joint attention behaviors. We validate the proposed system through series of interaction experiments. Experiment results showed that robot is able to identify objects that referenced by humans with the proposed communication system.

Nowadays, social media is getting more and more popular so that many people choose to absorb the knowledge, share their moods, read news, listen to music, and appreciate the video on the Internet. The popular Chinese songs can be categorized according to their song style, their released decade, their singer, and so on. Currently, the song is always classified as a single category, such as inspiration, love, or family. However, when people listen to a song, they will have a different feeling according to their moods in the moment. This paper adopts the lyrics of the popular Chinese songs on the Internet as the experimental samples. Then, we classify the songs based on the natural language processing, ontology, Word2Vec, and fuzzy inference mechanism. The adopted natural language mechanism contains term comparison and term similarity to compute the different-category weights. Additionally, we also use predefined ontology, knowledge base, and rule base to classify the songs. Moreover, we also adopt the multilayer perceptron neural network with the backpropagation algorithm to train the data under a supervised learning. The learned results are better than the ones of the fuzzy inference mechanism. In the future, this study will enhance ontology, knowledge base, and rule base as well as enlarge the number of experimental samples to improve the performance. Finally, we will combine music appreciation with the robot to make children learn the knowledge more interesting.

In order to solve the overlap link constraint, trapped movement, and computational cost problem in current IK model, this paper proposes a new coupled spiking neural network (CSNN) model which is combined with artificial neural network (ANN). Several references of end of effector's movement will be generated as training model of ANN. Current joint positions and angle values, movement direction and distance will be the input data. Angular velocity of every joint will be the output data. However, ANN structure and number of references will be minimized. As an alternative, CSNN will be implemented, where one joint angle is represented by a coupled neurons interconnected to each others. CSNN has feedback input from the current condition of arm robot, and its output will be combined with ANN's output. CSNN interconnection will be optimized using steady state evolutionary algorithm with several epoch. The proposed model is implemented to simulate multiple link planar robot. The result shows the effectiveness of the proposed model which succeeded in several trajectory tests with minimum computational cost.

In recent years, the aging of the population has become an important social issue for Japan and other developed countries. Especially for elderly people who live alone, due to lack of guardians, lonely death and accidents caused senile dementia occur from time to time. In this paper, we developed a monitoring system for the elderly based on the idea of Information Structure Space. By collecting environmental data we can estimate behavior of the elderly in the room. In this way, if the system finds an exception, an alert can be sent at any time. For this purpose, we use the fuzzy spiking neural network to realize the estimation of human behavior. In the experiment, we collected four kinds of environmental sensor data from bathroom and toilet. The result shows the system can estimate human's behavior with 94% accuracy.

Recently, the expectation to teleoperated mobile robots has been increasing much in order to perform the monitoring in various scenes. However, there are many critical problems in the teleoperated mobile robots. In this paper, we discuss cooperative formation behavior of teleoperated multiple robots. Especially, we focus on an enclosing formation behavior of a target object. First, we define the problem setting of the enclosing formation behavior. In our method, the enclosing formation is divided by two strategies in order to reduce the search space of robot poses. Next, we introduce Batch Learning Growing Neural Gas (BL-GNG) in order to improve the learning convergence and reduce the user-designed parameters in GNG. BL-GNG uses an objective function based on Fuzzy C-means for improving the learning convergence. Furthermore, we apply two-layers BL-GNG to decide the positions of enclosing formation. Finally, we show several experimental results of the proposed method.

In this paper we build a stress management framework that aims to help professionals in the health care sector manage their occupational stress. Our system employs chatbots and robots to conduct conversation with individuals in order to derive a measure of stress using a Sense of Coherence model. The outputs of this model then drive a Peer Support model which selects and administers an intervention with the aim of reducing measured stress. Our results show that the conversation model and Sense of Coherence model that we develop are capable of measuring stress and can be used by the peer support model to successfully select appropriate support actions.

A smart home needs to be human-centric, where it tries to fulfill human needs given the devices it has. Various works are developed to provide homes with reasoning and planning capability to fulfill goals, but most do not support complex sequence of plans or require significant manual effort in devising subplans. This is further aggravated by the need to optimize conflicting personal goals. A solution is to solve the planning problem represented as constraint satisfaction problem (CSP). But CSP uses hard constraints and, thus, cannot handle optimization and partial goal fulfillment efficiently. This paper aims to extend this approach to weighted CSP. Knowledge representation to help in generating planning rules is also proposed, as well as methods to improve performances. Case studies show that the system can provide intelligent and complex plans from activities generated from semantic annotations of the devices, as well as optimization to maximize personal constraints' fulfillment. Note to Practitioners-Smart home should maximize the fulfillment of personal goals that are often conflicting. For example, it should try to fulfill as much as possible the requests made by both the mother and daughter who wants to watch TV but both having different channel preferences. That said, every person has a set of goals or constraints that they hope the smart home can fulfill. Therefore, human-centric system that automates the loosely coupled devices of the smart home to optimize the goals or constraints of individuals in the home is developed. Automated planning is done using converted services extracted from devices, where conversion is done using existing tools and concepts from Web technologies. Weighted constraint satisfaction that provides the declarative approach to cover large problem domain to realize the automated planner with optimization capability is proposed. Details to speed up planning through search space reduction are also given. Real-time case studies are run in a prototype smart home to demonstrate its applicability and intelligence, where every planning is performed under a maximum of 10 s. The vision of this paper is to be able to implement such system in a community, where devices everywhere can cooperate to ensure the well-being of the community.

Quantum-inspired computing is an emerging research area, which has significantly improved the capabilities of conventional algorithms. In general, quantum-inspired hopfield associative memory (QHAM) has demonstrated quantum information processing in neural structures. This has resulted in an exponential increase in storage capacity while explaining the extensive memory, and it has the potential to illustrate the dynamics of neurons in the human brain when viewed from quantum mechanics perspective although the application of QHAM is limited as an autoassociation. We introduce a quantum-inspired multidirectional associative memory (QMAM) with a one-shot learning model, and QMAM with a self-convergent iterative learning model (IQMAM) based on QHAM in this paper. The self-convergent iterative learning enables the network to progressively develop a resonance state, from inputs to outputs. The simulation experiments demonstrate the advantages of QMAM and IQMAM, especially the stability to recall reliability.

Fuzzy Markup Language (FML) presented by IEEE Computational Intelligence Society (CIS) has been an IEEE Standard since May 2016. It is an XML-based language for designer to easily construct the knowledge base and rule base of the developed fuzzy logic system. In this paper, we propose an FML-based linguistic classification agent and apply it to popular Chinese songs' classification in social media environment. In addition, the lyrics are retrieved from Youtube, Facebook or Google+, and then we adopt Natural Language Processing (NLP) mechanism to deal with the document preprocessing. First, the domain experts construct the classification ontology model and design related categories for the application domain. Moreover, the fuzzy concept sets are also adopted in the related categories. Then, the Chinese Knowledge Information Processing (CKIP) tool is utilized to deal with the Chinese documents of the songs. Finally, the FML-based knowledge base and rule base of the classification agent are constructed for inferring the related categories of the song. The Fujisoft robot PALRO receives the classified songs and plays the song for the desired users. Experimental results show the proposed classification agent can work correctly.

In this paper, we present a summarization agent based on Fuzzy Markup Language (FML) and its application for meeting schedule data analysis. The knowledge base and rule base of FML are constructed by referring the meeting schedule ontology of Research & Development (R&D) office in National University of Tainan (NUTN) from Jan. 2011 to Jul. 2015. We propose an intelligent agent to retrieve the meeting activities of R&D Office from open meeting schedule database of NUTN. There are three categories belonging to the R&D meeting schedule ontology, including an International Affairs division, an Academic Development division, and an Industry-Academia Collaboration division. In addition, we apply the Natural Language Processing (NLP) open API for Chinese text mining and document preprocessing. Finally, the proposed FML-based summarization agent is combined with the human-friendly robot partner PALRO, produced by Fujisoft incorporated, to construct a meeting summarization robot agent. Experimental results show that the proposed agent can work effectively.

Recently, the development and research of robot partners are actively being carried out, and expectation for human-friendly robot partner has been increasing. In the development of a robot, it is necessary to consider development of robots for various needs according to the purpose of usage. Therefore, we propose design templates for easy design of conversation contents. This paper is organized as follows. First, we explain the robot partner system which has modular structure. This modular structure makes it easier to develop the robot partner. Next, we discuss the interaction system for robot partner and the design template to be applied to it. Finally, we show experimental results of the effectiveness of the interaction content design templates, and discuss the applicability of the proposed system for improving robot partner system.

This paper presents a novel combination method for on-line topological map building and dynamic path planning. The proposed method consists of two main components: Bayesian Adaptive Resonance Associative Memory (Bayesian ARAM) and forward-backward propagation path planner. Bayesian ARAM incrementally clusters sensory information and generates topological map. The explored environment is described as a group of neurons (nodes) and edges. Each neuron (nodes) represents a distinct place and it is defined as multi-dimensional Gaussian distribution which does not require any prior knowledge of what a place is supposed to be to make it works in natural environment. The topological map is incrementally generated by Bayesian ARAM. The forward-backward propagation path planner consists of two process: forward propagation determines the possible path while backward propagation with neuron pruning eliminates inefficient neurons and determines the optimum pathway from current location to target location based on the generated map information. The effectiveness of our proposed method is validated by several standardized benchmark datasets.

Aging society in Japan can be a big serious problem. However, the number of caregivers is currently not enough, and it is not expected to sufficiently increase in future. Elderly care has been shifting from hospital care to community-based care and home care, but this can lead to raise the burden on their family members. Therefore, elderly people should take care of their heath in daily life in order to prevent mental and physical depression. This paper presents a daily exercise support system with a robot partner utilized as an exercise instructor. We discuss the robot human-robot interaction in terms of framing effect. The framing effect is an example of cognitive bias that influences someone's choice depending on whether it is presented as a positive thing or as a negative thing. Here, we implemented verbal communication contents base on positive frame or negative frame. Furthermore, we conducted a demonstration experiment to examine the effect of each expression on elderlies' motivation.

In this paper, we demonstrate the application of Fuzzy Markup Language (FML) to construct an FML-based Dynamic Assessment Agent (FDAA), and we present an FML-based Human–Machine Cooperative System (FHMCS) for the game of Go. The proposed FDAA comprises an intelligent decision-making and learning mechanism, an intelligent game bot, a proximal development agent, and an intelligent agent. The intelligent game bot is based on the open-source code of Facebook’s Darkforest, and it features a representational state transfer application programming interface mechanism. The proximal development agent contains a dynamic assessment mechanism, a GoSocket mechanism, and an FML engine with a fuzzy knowledge base and rule base. The intelligent agent contains a GoSocket engine and a summarization agent that is based on the estimated win rate, real-time simulation number, and matching degree of predicted moves. Additionally, the FML for player performance evaluation and linguistic descriptions for game results commentary are presented. We experimentally verify and validate the performance of the FDAA and variants of the FHMCS by testing five games in 2016 and 60 games of Google’s Master Go, a new version of the AlphaGo program, in January 2017. The experimental results demonstrate that the proposed FDAA can work effectively for Go applications.

This paper proposes a gait generation system based on human behavior using biological approach. Humans have different gaits with different levels of speed or step. The proposed gait generator is able to generate the walking transition when the speed and the step length are changing dynamically. Neuron inter-connection structures as the locomotion model are formed. We apply evolutionary computation for each level of walking optimization. In locomotion generator, one joint angle is represented by two coupled neurons. Synaptic weights connected between ten motor neurons represent five joint angles and their gain values required to be optimized during several levels of speed. The optimized walking patterns are combined for acquiring dynamic relationship in one gait generator by using supervised multilayer perceptron (MLP) learning system. This gait generator uses optimized MLP weight parameters to generate synaptic weights transferred to locomotion generator depending on the desired walking speed. In order to prove the effectiveness of the model, we implemented it in computer simulation and in simple humanoid robot. The walking transitions depending on the changes in the walking speed are also shown. The smoothness of walking transition increased compared to previous researches.

Recently, the number of elderly people living alone is increasing and has become a serious problem in Japan. On the other hand, the stability of both social rhythm and biological rhythm is very important for extension of healthy life expectancy. It is difficult for elderly people to understand the current stability of social rhythm and biological rhythm in daily life. Therefore, we propose several visualization and management systems based on daily life monitoring. First, we explain a measurement method of human daily life logs. Next, we show a social rhythm management support system based on three different types of classification methods, such as (1) ICF (International Classification of Functioning, Disability and Health), (2) NHK (Nippon Hoso Kyokai) Activity classification, and (3) Social Rhythm Metric (SRM). Finally, we discuss the effectiveness of the proposed methods and future works.

This paper proposes a real-time feature extraction and segmentation method for a 3D point cloud. First of all, we apply Growing Neural Gas with Utility (GNG-U) to the point cloud for learning a topological structure. However, the standard GNG-U cannot learn the topological structure of 3D space environment and color information simultaneously. To this end, we then modify the GNG-U algorithm by using a weight vector. we propose a surface feature extraction and segmentation method by efficiently utilizing the topological structure. Our segmentation method is based on a region growing method whose similarity value uses the inner value of two normal vectors connected by the topological structure. We show experimental results of the proposed method and discuss the effectiveness of the proposed method.

This paper presents Bézier curve based passive neural control applied in bio-inspired locomotion in order to decrease the computational cost implemented for 4 legged animal robot which has 3 joints in each leg. Neural oscillator model is applied for generating the walking pattern in bio-inspired locomotion. Bézier curve based optimization represents passive neural control supported by evolutionary algorithm for representing the relationship equation between neuron signal and reference joint signal. Passive neural control is implemented in order to reduce the neuron complexity in neuro-based locomotion by controlling 3 joints with one signal without decreasing the performance both in walking pattern and in its stability level, whereas one leg is represented by one motor neuron. Therefore, the 4 legged robot is controlled by 4 motor neurons which have feedback connection with ground and inertial sensor. In order to prove the effectiveness, we implemented the model in computer simulation and in a small 4 legged robot. This model can decrease the computational cost so it is possible to apply the model in either animal or humanoid robot with low frequency processor.

This paper describes an air-coupled ultrasound system by using DIO-2000. The system is aimed to use for inner muscle evaluation in rehabilitation process. The system evaluates inner muscle with low constrain than conventional method which measured by using MR image, X-ray CT and contacted ultrasound system. Our air-coupled ultrasound system measures transmitted ultrasound wave with very low power through human abdominal region by employing a pulsar-receiver with high sensitive preamplifier, and wave detection method based on fuzzy inference finds transmitted wave from noisy wave. The fuzzy inference is derived from characteristics of transmitted wave. In the experiment, we evaluate the accuracies of wave detection method for human body.

An accurate and noninvasive stress assessment from human physiology is a strenuous task. In this paper, a pattern recognition system to learn complex correlates between heart rate variability (HRV) features and salivary stress biomarkers is proposed. Using the Trier social stress test, heart rate and salivary measurements were obtained from volunteers under varying levels of stress induction. Measurements of salivary alpha-amylase and cortisol were used as objective measures of stress, and were correlated with the HRV features using fuzzy ARTMAP (FAM). In improving the predictive ability of the ARTMAPs, techniques, such as genetic algorithms for parameter optimization and voting ensembles, were employed. The ensemble of FAMs can be used for predicting stress responses of salivary alpha-amylase or cortisol using heart rate measurements as the input. Using alpha-amylase as the stress indicator, the ensemble was able to classify stress from heart rate features with 75% accuracy, and 80% accuracy when cortisol was used.

This paper proposes the control system for 3-D locomotion of a humanoid biped robot based on a biological approach. The muscular system in the human body and the neural oscillator for generating locomotion signals are adapted in this paper. We extend the neuro-locomotion system for modeling a multiple neuron system, where motoric neurons represent the muscular system and sensoric neurons represent the sensor system inside the human body. The output signals from coupled neurons representing the angle joint level are controlled by gain neurons that represent the energy burst for driving the joint in each motor. The direction and the length of step in robot locomotion can be adjusted by command neurons. In order to form the locomotion pattern, we apply multiobjective evolutionary computation to solve the multiobjective problem when optimizing synapse weights between the motoric neurons. We use recurrent neural network (RNN) for the stabilization system required for supporting locomotion. RNN generates a dynamic weight synapse value between the sensoric neuron and the motoric neuron. The effectiveness of our system is demonstrated in open dynamic engine computer simulation and in a real robot application that has 12 degrees of freedom (DoFs) in legs and four DoFs in hands.

Human-like-motion performed by robots can have a contribution to exert a strong influence on human-robot interaction, because bodily expressions convey important and effective information. If the robots could adapt the features of human behavior to their motions and skills, the communication would become more smooth and natural. In this paper, we develop a posture measurement system for a robot imitation using a 3D image sensor. This paper proposes a method of robot posture generation based on a steady-state genetic algorithm (SSGA). SSGA is one of evolutionary optimization methods using selection, mutation, and crossover operators. Since SSGA is a simplified model, it is easy to implement into a real-time processing. Furthermore, we apply a continuous model of generation for an adaptive search in dynamical environment.

One of the problems in humanoid locomotion generation is energy efficiency. This paper proposes a method for energy efficiency optimization in simple humanoid robot locomotion using single objective genetic algorithm. With the aim to produce walking trajectory system using minimum energy and good stabilization, torque and oscillation analysis are required to calculate the stabilization. The number of desired outputs in this system is 4 parameters and the number of inputs is 9 parameters. We used neural network with back propagation learning mechanism to realize the relationship between input and output data as well as producing fitness function for genetic algorithm. The trajectory system has 2 trajectory equations, which is pelvis trajectory and ankle trajectory. Ankle trajectory is formed from circle function in Cartesian coordinate space and pelvis trajectory is formed from third order polynomial equation. Both of them are influenced by inclination of robot body. In the experiment, we apply this system using Bioloid robot with inertial sensor already installed. The experimental results show the analysis of energy by observing the torque resulted by servomotor in each joint. We observe that using this system, the torque value resulted by servomotors was decreased and has good stabilization.

In this paper, we propose locomotion generation for a mobile robot. Legged robot can walk in various complex terrains such as stairs as well as in flat environment. However, setting its behaviour to adapt to various environments in advance is very difficult. The robot can mimic the movement of organisms based on computational intelligence. In this study, we apply spiking neural network, which can take into account the transition of temporal information between the neurons. More specifically, the motion patterns are generated by applying a spiking neural network trained by Hebbian learning and evolution strategy, by using data provided by the physics engine measuring the distance walked by the robot and applied the motion patterns to real robot. Simulation was conducted to confirm the proposed technique.

In this study, we aim to develop a system for improving daily lives of elderly people to ensure health. In order to realize an enriched life style among elderly people, daily health care is important. Therefore, we have proposed a system where robot partners will assist in exercising activity among elderly people. This paper proposes a method of relational trust modeling based on reinforcement learning. We apply a concept of relational trust defined as the expectation that a person is disposed to act in a trustworthy manner toward "me," no matter what the person does to others. In the experiment, we discuss the effectiveness of relational trust for robot communication.

In this paper we present synaptic weight optimization for dynamic locomotion in four-legged robot (cat, dog) based on neural oscillator. We investigate the muscular structure to design the relationship for both extensor neuron and flexor neuron. The robot has 3 joints in each leg and each joint is represented by 2 neurons, extensor and flexor neuron. The robot has 4 main circular neurons as the server neuron and the other neurons are the client neurons. The server neurons generate the oscillator signal to the client neurons. The signal can be dynamically adjusted according to the environmental condition. Not only the synaptic network between the neurons, but the synaptic network between neurons and sensors was also designed to realize dynamical locomotion. Pressure sensor and inclination sensor were installed in the robot. The signal is influenced by ground reaction sensor and body inclination feedback. While the foot touches the ground, the sensory neuron sends the signal to the joint neuron. Negative signal will be sent to flexor neuron and positive signal will be sent to extensor neuron. To optimize the strength of weights in the synaptic neurons we apply the Nondominated Sorting Genetic Algorithm II (NSGA-II). The stability of torso body, the velocity, and the movement direction are the three objectives in the multi-objective NSGA-II. In the experiments, a computer simulation framework, the Open Dynamic Engine (ODE) is applied. The solution is evaluated based mainly on the moving distance of the robot. Experiments were conducted to confirm the proposed technique.

Health promotion support systems provide exercise program for elderly people to ensure their health. Some robot partners are used for the system as instructor. In the system, the instructor robot need to select an exercise that improves their health corresponding to the motivation of each person. Our system uses self-efficacy as the motivation index of elderly people for exercise. This paper proposes a self-efficacy estimation method to select more appropriate exercise. This method estimates the self-efficacy based on predicted exercise score acquired through communication and measured score obtained from distance distribution sensor.

In order to keep healthy health of elderly people, daily exercise is an important factor. Therefore, we have developed an exercise support system with robot partner to provide the daily exercise program. Furthermore, Human-like-motion can have a contribution to exert a strong influence on the person through the human-robot interaction, because bodily expressions convey important and effective information. If robots could adapt the features of human behavior to their motions and skills, the communication would become more smooth and natural. In this paper, we propose a learning structure for imitation learning.

This paper proposes a verbal conversation system for a robot partner using emotional model. The robot partner calculates its emotional state based on the utterance sentence of the human. Then, the robot partner can control its utterance sentence based on the emotional parameters. As a results, the robot partner can interact with human emotionally naturally. In this paper, we explain the three parts of the conversation system's structure. The first part is time dependent selection based on the database contents. In this mode, the robot tells timely important contents, for example schedules. The mood parameter is used to change the sentence in this mode. The second component is utterance flow learning to select the utterance contents. The robot selects utterance sentence based on the utterance flow information and using its mood value as well. The third component is sentence building based on predefined rules. The rules include personality model of the robot partner. In this paper, we use emotional parameters based on the human sentences to make a natural communication system. Finally, we show experimental results of the proposed method, and conclude the paper. The future research for improving the robot partner system is discussed as well.

In this paper we propose the development of EROS locomotion by using neural oscillator. We investigated muscular structure of human body for designing the neuron structure. Two motoric neurons, extensor neuron and flexor neuron, represent one structure of joint that generating the angle of joint. Sensoric neuron connection also designed for adapting the environment. Three kinds of sensor such as ground reaction sensor, tilt sensor, and angular velocity sensor are utilized for validate the proposed method. Evolutionary algorithm was used for optimizing synapse weight among motoric neuron, while recurrent neural network was used for the dynamical condition learning. The locomotion system of this research was shown using Open Dynamic Engine (ODE). The proposed method can generate locomotion pattern and its stability learning system improves the stability of locomotion. The proposed approach formed the walking locomotion that potentially can be developed to become adaptive locomotion.

Aphasia is one of the conditions of a higher brain function dysfunction. The aphasia decreases quality of life of patient with the dysfunction. Moreover, since patients with aphasia have handicap for talking, reading and understanding about rehabilitation program, the aphasia prevents progress of other rehabilitation. This paper describes a rehabilitation support system for patient with aphasia. The system employ tablet device as multimodal interface device. The system provides several tasks and measures their solution times and answers instead of speech therapist. And, the measured data send to cloud database to diagnose progress of that. In experiment, usability of our system was tested with six aphasia patients.

Recently, various types of wireless sensor network systems have been developed to realize daily care for elderly people living alone. Furthermore, visualization methods of life logs have been presented. However, it is important to integrate different types of data measured by each sensor node to estimate human states and behaviors. Therefore, we have proposed the concept of informationally structured space (ISS). This paper proposes a methodology to deal with data measured by sensor nodes in wireless sensor networks on ISS. First, we explain how to use ISS for wireless sensor networks. Next, we apply the proposed method to elderly care. We propose four different components such as (1) human localization by spiking neurons, (2) human movement transition probability, (3) redundant monitoring by simultaneous firing of sensor nodes, and (4) temporal life pattern extraction by Gaussian membership functions. Finally, we show several simulation results and discuss the effectiveness of the proposed method.

In this research, we propose a methodology for getting joint angles by Kinect sensor for rehabilitation evaluation support. We measure the motion of the arm of a patient with hemiplegia before and after the rehabilitation, and estimate the range of the motion by using genetic algorithm and neural network. The range after the rehabilitation is bigger than before the rehabilitation. Based on this result, our methodology is able to evaluate the change of the motion before and after the rehabilitation for patients with hemiplegia.

This paper discusses robot partner interaction based on Frankl's psychology to encourage a person act voluntarily. Recently, elderly people who live alone in a room is increased. But, it is hoped that the elderly people keep good health. To keep good health for elderly people, a person should act voluntarily and get involved in community events. In Frankl's psychology insists that the meaning of life is required from external stimuli. Frankl's psychology is defined 3 fields of value which are provided from external stimuli. A person can be found the meaning of life according to the values getting. Therefore, it is expected that a person would act voluntarily if a person find the meaning and value of life from external stimuli through interacting with a robot partner. In this research, we propose the robot interaction method based on Frankl's psychology. And we propose an external value estimation method based on the fields of value which estimates an external value from environmental change. Through human-robot interaction experiment based on Frankl's psychology, we verify that a human acts voluntarily by finding the external value.

Biological systems are said to learn from both intrinsic and extrinsic motivations. Extrinsic motivations, largely based on environmental conditions, have been well explored by Reinforcement Learning (RL) methods. Less explored, and more interesting in our opinion, are the possible intrinsic motivations that may drive a learning agent. In this paper we explore such a possibility. We develop a novel intrinsic motivation model which is based on the well known Yerkes and Dodson stress curve theory and the biological principles associated with stress. We use a stress feedback loop to affect the agent's memory capacity for retrieval. The stress and memory signals are fed into a fuzzy logic system which decides upon the best action for the agent to perform against the current best action policy. Our simulated results show that our model significantly improves upon agent learning performance and stability when objectively compared against existing state-of-the-art RL approaches in non-stationary environments and can effectively deal with significantly larger problem domains.

This paper discusses a robot partner system for natural communication using emotional models. In our daily life, robot partners should have an emotional model in order to co-exist and to realize natural communication with people. In this paper, we propose several emotional models for human-robot interaction based on computational intelligence. First we discuss the importance of emotion and its functions in the social interaction. Next, we propose an emotional model based on emotion, feeling, and mood. Furthermore, we use the emotional model as a method for communication system, and also, we discuss Frankl's psychology as the basis of communication. Finally, we show several experimental results of the proposed method, and discuss the utterance systems for a robot partner.

This paper proposes a method of feature extraction from 3D point clouds for informationally structured space including sensor networks and robot partners for co-existing with people. The informationally structured space realizes the quick update and access of valuable and useful information for both people and robots on real and virtual environments. Our method is based on Hierarchical Growing Neural Gas (HGNG). This method is one of self-organizing neural network based on unsupervised learning First, we propose 3D map building method using Kinect in order to acquire the 3D point clouds. Next, we propose the method of the feature extracting method based on HGNG. Finally, we show experimental results of the proposed method and discuss the effectiveness of the proposed method.

Recently, language education has a great demand from elementary school to adults. Robots are used as teaching assistants in Robot-Assisted Language Learning. It is very effective to use robots for language education. However, the robots may have some problems. One of the problems is to get bored when interacting with robots. This paper deals with this issue by using a method based on social cognitive theory. We discuss the role of robots based on mutual learning in language education. Next, we explain the concept of self-efficacy for evaluating the learning condition of robots. We propose a method to express self-efficacy using fuzzy control. The essence of the proposed method is to adapt to human's state. The experimental results show the effectiveness of the proposed method for long-term communication between a human and a robot.

This paper deals with the information extraction of daily life log measured by smart phone sensors. Two types of neural computing are applied for estimating the human activities based on the time series of the measured data. Acceleration, angular velocity, and movement distance are measured by the smart phone sensors and stored as the entries of the daily life log together with the activity information and timestamp. First, growing neural gas performs clustering on the data. Then, spiking neural network is applied to estimate the activity. Experiments are performed for verifying the effectiveness of the proposed method.

This paper proposes a conversation system based on multimodal perception for verbal communication between a human and a robot partner using various types of sensors. First, we describe the control structure of the robot partner and explain the architecture of the robot system. Next, evolutionary robot vision is applied to human and object detection. Next, a conversation system based on information ally structured space is proposed. Furthermore, we propose a method of conversation learning based on the flow of human utterance patterns and its related perceptual information. Finally, we show experimental results of the proposed method, and discuss the future direction on this research.

Recent technology and market trends have demanded the significant need for feasible solutions to video/camera systems and analytics. This paper provides a comprehensive account on theory and application of intelligent video systems and analytics. It highlights the video system architectures, tasks, and related analytic methods. It clearly demonstrates that the importance of the role that intelligent video systems and analytics play can be found in a variety of domains such as transportation and surveillance. Research directions are outlined with a focus on what is essential to achieve the goals of intelligent video systems and analytics.

This paper proposes a localization method using multiresolution maps for the navigation of multiple mobile robots based on formation behaviors. The remote control of multiple mobile robots is one the most important tasks in robotics to realize distributed remote monitoring in unknown and/or dynamic environments. However, it is very difficult for a human operator to control multiple mobile robots separately at the same time. Therefore, autonomous formation behaviors of multiple robots are required to reduce mental and physical loads of the human operator. If each mobile robot can estimate the self-position or relative position in a group, it is easier for multiple mobile robots to realize formation behaviors. First, we propose a method of simultaneous localization and mapping based on a grid approach. Next, we explain how to share the build map among multiple mobile robots, and propose a self-localization method based on multiresolution maps. Furthermore, we explain the formation behaviors of multiple mobile robots. Finally, we show several experimental results, and discuss the effectiveness of the proposed method.

This paper proposes a method for 3D modeling of environments used to perform teleoperation of a mobile robot. Recently, the expectation to tele-operated mobile robots has been increasing much in order to perform a monitoring in various scenes. However, there are many critical problems in tele-operated systems. Especially, we must expand visual range from a robot, the usability of human interface, and intention sharing between the robot and operator. First, we discuss information visualization for human-friendly tele-operation. Next, we propose a tele-operating system based on multi-resolution map. Finally, we propose a method of 3D modeling using Microsoft Kinect sensor, and show several experimental results of the proposed method.

Recently, the need of robot partners is increasing. Such robots should have an emotional model in order to co-exist and to realize the natural communication with people. In the communication, nonverbal communication and emotional expression based on emotional model are very important for robot partners. Moreover, facial and gestural expression should be adaptive to a user of the robot. Therefore, we propose a design support system of arm gestural and facial expression of robot partners based on interactive evolutionary computation and Laban features. Next, we conduct several experiments of the proposed method, and discuss the effectiveness of the proposed method.

This paper proposes a method for constituting the formation of a multi-robot system according to dynamically changing environments. First, we apply a method of multi-objective behavior coordination for integrating behavior outputs from the fuzzy control for collision avoidance and target tracing. Second, we apply a spring model to calculate the temporary target position of each robot for the formation behavior. Third, we discuss multi-robot behaviors based on the concept of coupling. The tight coupling is realized by the spring model while the loose coupling is realized by the individual decision making based on connection and disconnection with other robots. Furthermore, the proposed method is applied to the exploration in unknown environments. Finally, we discuss the effectiveness of the proposed method through several simulation results.

This paper deals with method of interactive categorization based on simultaneous localization and mapping for user support by a robot partner. We propose a method of getting information about unknown objects in the two-dimensional map. First, we explain a method of updating a robot location by a steady-state genetic algorithm. Second, we explain a map building method based on the topological approach by a growing neural network. Next, we explain a method of noise reduction in the map caused by moving objects. Furthermore, we propose an estimation method of human position and object areas by the communication from a robot partner to the human. Finally, we show several experimental results of the proposed method, and discuss the effectiveness on this research.

The 11 papers in this special section focus on intelligent video systems and analytics.

This paper proposes a multimodal communication method for human-friendly robot partners based on various types of sensors. First, we explain informationally structured space to extend the cognitive capabilities of robot partners based on environmental systems. Next, we discuss the suitable measurement range for recognition technologies of touch interface, voice recognition, human detection, gesture recognition, and others. Based on the suitable measurement ranges, we propose an integration method to estimate human behaviors based on the human detection using color image and 3-D distance information, and gesture recognition by the multilayered spiking neural network using the time series of human-hand positions. Furthermore, we propose a conversation system to realize the multimodal communication with a person. Finally, we show several experimental results of the proposed method, and discuss the future direction of this research.

This paper discusses human preference learning by robot partners through interaction with a person. We use a robot music player; miuro, and we focus on the music selection for providing the person with comfortable sound field. First, we propose a control architecture of miuro based on autonomous behavior mode, interactive behavior mode, and human control mode. Next, we propose a learning method of the relationship between human position and its corresponding music selection based on Q-learning. Furthermore, we proposed a similarity matrix to reduce the learning time of Q-learning. The experimental results show that the proposed method can learn the relationship between human position and its corresponding human preferable music.

This paper discusses measurement methods of human behaviors based on sensor network and human interaction of rehabilitation using robot partners. First, we explain robot partners and sensor networks for rehabilitation. Next, we apply a steady-state genetic algorithm to extract human motions from 3D distance image. Finally, we discuss the effectiveness of the proposed methods through several experimental results.

Recently, multi-robot systems have been discussed to realize a large size of distributed autonomous system. Furthermore, multi-robot systems have been applied to various problems such as autonomous guided vehicles, soccer robots, and search and rescue system by multi-robot. This paper proposes intelligent formation behavior for the multi-robot based on sensor fusion. First, we discuss multi-agent systems and wireless network technologies. Next, we explain the hardware specification of robot and tele-operated system and wireless communication. Finally, we show experimental results, and discuss the availability of intelligent formation behavior for multi-robot.

This paper discusses robot perception of an unexpected object for human friendly robots. A robot should be able to perceive an environment flexibly to realize an intelligent behavior. We focus on a perceptual system for human visual perception based on perceiving-acting cycle concept discussed in ecological psychology. We have proposed a perceptual system composed of the retinal model and the spiking-neural network to realize the concept of the perceiving-acting cycle. The proposed method is applied to a robot arm equipped with a 3D range camera. In this paper, we propose an integrated perceptual system for accuracy improvement of perception. Moreover, we propose a perceptual element to install to the integrated perceptual system that detects an unexpected posture or object by using 3D range camera. As experimental results, we show that the proposed method perceives the target dish accurately by an integration of different perceptual elements, and the robot recognizes an unexpected situation such as a fallen cup and a ball of paper.

This paper discusses the social learning of robot partners through interaction with a person. We use a robot music player; Miuro, and we focus on the music selection for providing the comfortable sound field for the person. First, we propose the control architecture of Miuro based on autonomous behavior mode, interactive behavior mode, and human control mode. Next, we propose a learning method of the relationship between human interaction and its corresponding reaction based on Boltzmann selection, adaptive reward function, and temperature control. The experimental results show that the proposed method can learn the relationship between human interaction and its corresponding behavior, even if the human intention is changed in the learning. Furthermore, the experimental results show that the proposed method can provide the person the preferable song as the comfortable sound field.

Recently, the need of users is changing to the efficient quality of functions with the sophisticated design and reasonable price. Furthermore, current users prefer to the personal customization of products. Accordingly, a design support system is useful and helpful for non-expert people to design products easily, but such non-expert people might take much time and load in the product design. Therefore, we proposed interactive design support system based on evolutionary computation, and applied the proposed method to the design of robot partners. However, it is very difficult to reflect human evaluation to the generation of the next design candidates. Therefore, we propose an estimation method of human evaluation using fuzzy inference in the interactive design support using the evolutionary computation. Furthermore, we use iPhone simulator to evaluate the human impression based on direct interaction with the designed robot partner. Finally, we discuss the effectiveness of the proposed system through several simulation and experimental results.

By using human speech information, different kinds of speaker and speech recognition systems have been developed for partner robots to efficiently cooperate with people in the daily life. For improving the recognition accuracy and robustness, a two-stage pattern matching algorithm for speaker recognition system of partner robots is proposed. In the first matching stage, by using fuzzy c-means and declustering in vector quantization(VQ) method, the recognition performance with limited training data is improved. For avoiding the phenomenon of similar cepstral features by different speakers, with three additional speech features, the second stage is designed to rematch the similar recognition results of the first stage. In order to evaluate the proposed structure, some experiments have implemented on a public database ELSDSR and an speech owners database for partner robots. The results verified the proposed method obtained more accurate recognition results with strong robustness.

This paper proposes a human localization method in informationally structured space based on sensor network First, we explain informationally structured space, robot partners, and sensor networks developed in this study. Next, we apply a fuzzy spiking neural network to extract a person from the measured data by the sensor network. Furthermore, we propose a learning method of fuzzy spiking neural network based on the time series of measured data. Finally, we discuss the effectiveness of the proposed methods through experimental results in a living room.

This paper proposes a measurement method of human position based on sensor network First, we explain informationally structured space, robot partners, and sensor networks developed in this study. Next, we discuss the applicability of the sensor network and robot partners for human interaction. Next, we apply a steady-state genetic algorithm using template matching to extract a person in 3D distance image based on differential extraction. Finally, we discuss the effectiveness of the proposed methods through several experimental results.

Fast and flexible walking is necessary for hu-manoid robots in the Robocup soccer competition. Instability is one of the major defects in humanoid robots. Recently, various methods on the stability and reliability of humanoid robots have been actively studied. We propose a new fuzzy-logic control scheme that would enable the robot to realize flexible walking or turning with high standard of stability by restricting the step length and inclining the body of robot to an appropriate extent. In this paper, a stabilization algorithm is proposed using the balance condition of the robot, which is measured using accelerometer sensors during standing, walking, and turning movement are estimated from these data. From this information the robot selects the proper motion pattern effectively. In order to generate the proper reaction under various the body of robot situations, a fuzzy algorithm is applied in finding the proper angle of the joint. The performance of the proposed algorithm is verified by walking, turning tap and ball kicking movement experiments on a 18-DOFs humanoid robot, called EFuRIO.

This paper proposes a method of 3D topological reconstruction for informationally structured space including sensor networks and robot partners for co-existing with people. The informationally structured space realizes the quick update and access of valuable and useful information for both people and robots on real and virtual environments. In this paper, we use distance information and color information measured by 3D distance image sensor and CMOS camera for 3D topological reconstruction. First, we propose an extraction method of objects from the background image based on Hough transform as preprocessing. Next, we propose a method of 3D topological reconstruction based on growing neural gas to construct informationally structured space. Finally, we show experimental results of the proposed method and discuss the effectiveness of the proposed method.

Recently, the need of service robots is increasing, and the performance of intelligent technology for perceiving objects and people should be improved. We focus on a service robot system for clearing a table as a human-friendly task. The task is composed of (1) target object detection, (2) the estimation of position and posture of the target objects, and (3) the clearing of the target object. Therefore, this paper proposes a method of estimating the position and posture of the target object based on Hough transform and steady-state genetic algorithm. Next, we define the perceptual index to evaluate the state of perception to specify the position and posture of target objects. Furthermore, we discuss the availability of the proposed perceptual index through preliminary experimental results.

This paper discusses a integrated perceptual system for intelligent control of a service robot. The robot should be able to perceive the environment flexibly to realize intelligent behavior. We focus on the perceptual system based on the perceiving-acting cycle discussed in ecological psychology. We propose a perceptual system composed of the retinal model and the spiking-neural network to realize concept of the perceiving-acting cycle. The proposed method is applied to the robot arm equipped with a 3D-range camera. This proposed method features the robot detects the invariant information of a dish, for example the contrast of a distance Information or a luminance information. As experiments, we discuss about the effectiveness of the proposed method by comparing different input.

Recently, multi-agent systems have been discussed to realize a large size of distributed autonomous system. This paper proposes an intelligent control method for formation behaviors of multi-robot. First of all, we discuss the current state of researches on formation behaviors in multi-robot. Next, we propose a multi-objective behavior coordination to realize formation behavior based on the integration of the intelligent control from the local viewpoint of individual intelligence and the spring model from the global viewpoint of collective intelligence. Next, we propose a self-adaptation method in complicated environments. Finally, we discuss the effectiveness of the proposed method through computer simulation results.

This paper discusses a perceptual system for an intelligent service robot based on the Perceiving-Acting Cycle. Recently, various intelligent robots work in real environments such as public facilities, commerce facilities and houses. We are developing an intelligent service robot. The task of this robot is to clear the table in a restaurant. In this environment, the robot must perceive the necessary information from various information to take a flexible action like a human. In this study, we focus on the perceptual system based on perceiving-acting cycle discussed in ecological psychology. First, we propose a retinal model for a 3D-range camera based on human retinal structure, and the information extraction method using a spiking neural network based on perceiving-acting cycle. Next, we apply the proposed method for a task of clearing the table. As an experimental result, we show the proposed method can detect a dish in dynamic environment. We discuss the efficiency of our proposed method.

Human interaction based on conversation and gestures is very important to realize the natural communication. This paper proposes a conversation system composed of topic selection module, conversation control module and utterance selection module. First, we apply a Bayesian network for the topic selection, and Boltzmann selection for the control of conversation. We apply term frequency inverse document frequency for representing the features of a document by a weight vector of terms used in the document. The experimental results show that the proposed method can select topics according to the perceptual information and human interaction.

This paper proposes an intelligent monitoring system based on emotional model in sensor networks. Emotional models are very useful to understand human behaviors. First, we explain the recent works on emotional models, and discuss the applicability of emotional models to real world problems. Next, we propose an emotional model composed of emotion, feeling, and mood. These three components are coupling, but their time scales are different. Next, we explain perceptual system based on image processing and the monitoring system based on the proposed emotional model. Finally, we discuss the effectiveness of the proposed method through several experimental results.

This paper discusses a integrated perceptual system for intelligent control of a service robot. The robot should be able to perceive the environment flexibly to realize intelligent behaviors. We focus on the perceptual system based on the perceiving-acting cycle discussed in ecological psychology. We propose a perceptual system composed of the retinal model and the spiking-neural network to realize concept of the perceiving-acting cycle. The proposed method is applied to the robot arm equipped with a 3D-range camera. This proposed method features the robot detects the invariant information of a dish, for example the contrast of a distance Information or a luminance information. As experimental results, We show that the integrated perceptual system can adapt effectively in the dynamic environment.

The understanding of human motions captured in image sequences pose two main difficulties which are often regarded as computationally ill-defined: 1) modelling the uncertainty in the training data, and 2) constructing a generic activity representation that can describe simple actions as well as complicated tasks that are performed by different humans. In this paper, these problems are addressed from a direction which utilises the concept of fuzzy qualitative robot kinematics [9]. First of all, the training data representing a typical activity is acquired by tracking the human anatomical landmarks in an image sequences. Then, the uncertainty arise when the limitations of the tracking algorithm are handled by transforming the continuous training data into a set of discrete symbolic representations - qualitative states in a quantisation process. Finally, in order to construct a template that is regarded as a combination ordered sequence of all body segments movements, robot kinematics, a well-defined solution to describe the resulting motion of rigid bodies that form the robot, has been employed. We defined these activity templates as qualitative normalised templates, a manifold trajectory of unique state transition patterns in the quantity space. Experimental results and a comparison with the hidden Markov models have demonstrated that the proposed method is very encouraging and shown a better successful recognition rate on the two available motion databases.

Recently multi-agent systems have been discussed to realize a large size of distributed autonomous system. This paper proposes an intelligent control of multiple partner robots as one of multi-agent systems. First of all, we discuss the current state of researches on the multi-agent systems. Next, to realize a formation behavior, we propose a multi-objective behavior coordination to realize formation behavior based on the integration of the intelligent control from the local viewpoint of individual intelligence and the spring model from the global viewpoint of collective intelligence. Finally, we discuss the effectiveness of the proposed method through several computer simulation results.

Recently, various types of human-friendly robot have been developed. Such robots should perform voice recognition, gesture recognition, and others. This paper discusses the learning capability of a human gesture recognition method based on computational intelligence. The proposed method is composed of image processing for human face and hand detection based on a steady-state genetic algorithm, an extraction method for human hand motion based on a fuzzy spiking neural network, and an unsupervised classification method for human hand motion based on a self-organizing map. We show several experimental results and discuss their effectiveness.

Recently multi-agent systems have been discussed to realize a large size of distributed autonomous system. This paper proposes an intelligent control method of multiple partner robots as one of multi-agent systems. First of all, we discuss the current state of researches on the multi-agent systems. Next, to realize a formation behavior, we propose a multi-objective behavior coordination to realize formation behavior based on the integration of the intelligent control from the local viewpoint of individual intelligence and the spring model from the global viewpoint of collective intelligence. Finally, we discuss the effectiveness of the proposed method through several computer simulation results.

This paper discusses a perceptual system for a partner robot from the viewpoint of human visual perception. Recently, various types of robots equip various types of sensors for perceiving the environment. However, the robot must perceive the necessary information from too much information to take a flexible action like a human. In this study, we emphasize the importance of human vision for the robot to realize perception and action flexibility. Especially we focus on the perceptual system based on perceiving-acting cycle discussed in ecological psychology. First, we propose a retinal model for a laser range finder based on human retinal structure, and the information extraction method using a spiking neural network based on perceiving-acting cycle. Next, we apply the proposed method for a human tracing task in a dynamic environment. As an experimental result, we show the robot can directly perceive the necessary information by the attention mechanism for the flexible perception according to spatiotemporal context based on the spiking neural network.

This paper proposes a steady-state genetic algorithm for trajectory generation used in the imitation of a partner robot interacting with a human. Various types of genetic algorithms have been applied for the trajectory generation of robot manipulators. In this paper, we propose a trajectory generation method for the partner robot by a steady-state genetic algorithm based on the human motions pattern, and compare the proposed method with its related methods. Finally, we show experimental results of trajectory generation through interaction with a human.

This paper discusses the role of evolutionary computation in visual perception for partner robots. The search of evolutionary computation has many analogies with human visual search. First of all, we discuss the analogies between the evolutionary search and human visual search. Next, we propose the concept of evolutionary robot vision, and a human tracking method based on the evolutionary robot vision. The proposed method is composed of human detection and the update of human tracking positions. Finally, we show experimental results of the human tracking to discuss the effectiveness of our proposed method.

Recently, various types of human-friendly robot have been developed. Such robots should perform voice recognition, gesture recognition, and others. This paper discusses the learning capability of a human gesture recognition method based on computational intelligence. The proposed method is composed of image processing for human face and hand detection based on a steady-state genetic algorithm, an extraction method for human hand motion based on a fuzzy spiking neural network, and an unsupervised classification method for human hand motion based on a self- organizing map. We show several experimental results and discuss their effectiveness.

Recently, various types of artificial neural networks are applied for behavioral learning of mobile robots in unknown and dynamic environments. In this research, the behavioral learning method based on a spiking neural networks for multiple mobile robots are proposed. The robots learn the forward relationship from sensory inputs to motor outputs. However, the behavioral leaning capability of the robots depends strongly on the network structure and the environments. Therefore, we use a parallel genetic algorithm for updating the network structure through the interaction among robots suitable to the environment. Finally, the effectiveness of the proposed method is discussed through experimental results on behavioral learning for collision avoidance.

This paper discusses the advantage and disadvantage of evolutionary robot vision and particle swarm optimization for multiple human heads tracking. Evolutionary robot vision combines the technologies of the evolutionary computation and robot vision. Both of evolutionary computation and particle swarm optimization can perform the multiple human heads tracking well for feasible solution in a dynamic movement. This paper compares their performance. Finally, the proposed method is applied to a partner robot, and we discuss the effectiveness of the multiple human heads tracking in the natural communication with humans.

This paper discusses structured learning for the prediction-based control of perceptual modules of partner robots. A partner robot should classify and predict human behavior patterns to control perceptual modules for natural communication with a human. Therefore we proposed a prediction-based perceptual system. The proposed system has three main functions; (1) the clustering of perceptual information (the extraction of spatial patterns), (2) the prediction of transition among the clusters (the extraction of temporal patterns), and (3) selection of perceptual modules (the control of sampling intervals). Finally, we show experimental results on the interaction with a human to discuss the effectiveness of our proposed method.

This paper discusses the advantage and disadvantage of evolutionary robot vision for tracking multiple humans. Evolutionary robot vision combines the technologies of the evolutionary computation and robot vision. The evolutionary computation can search for feasible solution in a dynamic environment. This paper shows its performance. The multi-scale feature extraction, working memory area and evolutionary algorithm (SSGA) were proposed. The proposed method can perform the multiple human heads tracking and also the computational cost can be reduced. Finally, the proposed method is applied to a partner robot, and we discuss the effectiveness of the multiple humans tracking in the natural communication with humans.

This paper discusses the role of prediction in the structured learning for the prediction-based perceptual system of partner robots. The perceptual system for a partner robot must perform many functions with many parameters for extracting necessary perceptual information from the viewpoint of embodiment. The robot requires the learnability and adaptability to regulate these parameters by itself, because the parameters cannot be pre-defined and fixed in communication with human. Predictive capability is also required to the robot. The robot can use each function in the perceptual system by reflecting the prediction result efficiently. Therefore we propose the prediction-based perceptual system. Each function in the proposed system enhances the learning of other functions by regulating parameters based on the concept of structured learning. Finally, we show experimental results on the interaction with a human to discuss the effectiveness of our proposed method.

This paper discusses the visual perception for natural communication between a partner robot and a human. The prediction is very important to reduce the computational cost and to extract the perceptual information for the natural communication with a human in the future. Therefore we propose a prediction-based control of visual perception based on spiking neurons. The proposed method is composed of four layers: the input layer, clustering layer, prediction layer, and perceptual module selection layer. Next, we propose a competitive learning method to perform the clustering of human behavior patterns. Furthermore, the robot select perceptual modules used in the next perception according to the predicted perceptual mode. The results of prediction are evaluated based on the Gaussian membership function. Furthermore, we show experimental results of the communication between a partner robot and a human based on our proposal method.

This paper discusses the emotional learning for natural communication of a partner robot. The robot should communicate with a human according to the facing environment. In this paper, we propose a perceptual system based on the emotional model. The emotional model updates the emotional states according to the sensed environmental information and human reactions. Next, we propose a learning method based on the emotional model. Finally, we show several experimental results of the communication and the emotional learning between the partner robot and the human.

Imitation is a powerful tool for behavior learning and human communication. Basically, imitative learning is composed of model observation and model reproduction. This paper applies a spiking neural network and self-organizing map for model observation, and modular fuzzy neural networks and a steady-state genetic algorithm for model reproduction. The proposed method is applied for a partner robot interacting with a human. Experimental results show that the proposed method enables a robot to learn behaviors through imitation and can interact with a human efficiently.

This paper proposes a perceptual system for communication of a partner robot based on computational intelligence. Basically, communication is restricted by the environment. Therefore, the robot should perceive the environment the robot is facing and communicates with human naturally. From this point of view, we propose the vision-based perceptual system for environmental learning. We also propose the learning method using bidirectional spiking neural networks for learning the relationship among the linguistic terms, gestures, and objects build on the environmental state. Furthermore, we show experimental results of a partner robot, MOBiMac

This paper discusses prediction of human behavior patterns for natural communication between a partner robot and a human. The prediction is very important to extract the perceptual information for the natural communication with a human in the future. Therefore we propose a prediction-based perceptual system based on spiking neurons. The proposed method is composed of four layers: the input layer, clustering layer, prediction layer, and perceptual module selection layer. In the clustering layer, an unsupervised learning method is used to perform the clustering of human behavior patterns. We use unsupervised learning because the human behavior patterns to be paid attention change by the other and the situation in communication. Furthermore, we show experimental results of the communication between a partner robot and a human based on our proposed method

This paper proposes swarm intelligence for a perceptual system of a partner robot. The robot requires the capability of visual perception to interact with a human. Basically, a robot should perform moving object extraction and clustering for visual perception used in the interaction with a human. In this paper, we propose a total system for human classification for a partner robot by using particle swarm optimization, k-means, self organizing maps and back propagation. The experimental results show that the partner robot can perform the human clustering and classification

This paper proposes a steady-state genetic algorithm for self-localization and map building for illuminance measurement of a mobile robot. The map is represented by 2 dimensional discrete cell space. According to the measured distance by laser range finder, the map is updated sequentially. When the difference between the measured distance and the map data is large, a steady-state genetic algorithm corrects the self-location. Finally we show computer simulation and experimental results of the proposed method.

This paper discusses communication between a partner robot and human based on visual tracking, and imitative learning for the partner robot. In this paper, we propose imitative learning and communication with human based on spiking neural network, self-organizing map, and steady-state genetic algorithm. Furthermore, we show experimental results of the partner robot based on imitation.

This paper proposes a human recognition method of a partner robot for natural communication with human. Basically, human recognition is performed by using various types of information. In this paper, we use the color image of human face and pattern of conversation with the human. The proposed method is composed of k-means algorithm, spiking neural network, self-organizing map, and steady-state genetic algorithm. Furthermore, we show experimental results of the partner robot based on the proposed method.

The aim of This work is to develop partner robots that can obtain and accumulate human-friendly behaviors. To realize it, we use a concept of structured learning which emphasizes the importance of an interactive learning of several modules through interaction with its environment. In a proposed method, a robot obtains hand-to-hand behavior by using an interactive evolutionary computation based on human evaluations estimated by fuzzy state-value functions. Moreover, a self-organizing map is used for clustering human hand positions. A state-value function and a knowledge database are assigned to each clustered positions. Furthermore, the best trajectory is stored in the knowledge database to reuse it in the same situation. Some experimental results show the effectiveness of the proposed method.

This paper introduces a learning method, structured learning, for partner robots interacting with a human. Soft computing techniques such as fuzzy, neural, and evolutionary computing methods are applied for learning the interrelation between a robot and a human. We show several experimental results of two types of partner robots.

This paper proposes a method for visual perception for a partner robot interacting with a human. A robot with a physical body should extract information by using prediction based on the dynamics of its environment, because the computational cost can be reduced, imitation is a powerful tool for gestural interaction between children and for teaching behaviors to children by parent. Furthermore, others' action can be a hint for obtaining a new behavior that might not be the same as the original action. This paper proposes a visual perception method for a partner robot based on the interactive teaching mechanism of a human teacher. The proposed method is composed of a spiking neural network, a self-organizing map, a steady-state genetic algorithm, and softmax action selection strategy. Furthermore, we discuss the interactive learning of a human and a partner robot based on the proposed method through several experiment results.

This paper proposes an on-line learning method for a partner robot. First, the concept of perceiving-acting cycle is applied for learning the relationship between perception and action of a partner robot interacting with its environment. Next, we propose a spiking neural network for learning collision avoiding behavior. The robot learns the forward relationship from sensory inputs to motor outputs as well as the predictive relationship from motor outputs to the sensory inputs. Experimental results show that the robot can learn embodied actions restricted by its physical body.

This paper proposes a method for generating behaviors based on imitation of a partner robot interacting with a human. First of all, we discuss the role of imitation, and explain the method for imitative behavior generation of the robot based on computational intelligence. The robot searches for a human by using a CCD camera. A human hand motion pattern is extracted from a series of images taken from the CCD camera. Next, the position sequence of the extracted human hand is used as inputs to a spiking neural network in order to recognize it as a gesture. Furthermore, the trajectory for a behavior is generated and updated by a steady-state genetic algorithm based on human motions. Furthermore, a self-organizing map is used for clustering human hand motion patterns as gestures. Finally, we show several experimental results of imitative behavior generation through interaction with a human.

This paper deals with multi-objective behavior coordination of multiple robots interacting with a quasi-ecosystem which is composed of insects and plants. In this ecosystem, there co-exist plants and insects according to specific reproduction rules. In general, the inhabiting area of each species is localized owing to geographical, climatic, and ecological factors. This indicates the population density of each species in one area is different from another according to local environmental conditions. In this study. multiple robots are introduced in order to maintain the ecosystem. Each robot takes actions based on multi-objective behavior coordination integrating several action outputs. However, the robot must select its suitable area in order to adapt to the current state of the quasi-ecosystem that might change dynamically. In this paper, we discuss target selection for insect removing and plant reaping behaviors through several computer simulations in a dynamically changing environment.

This paper is concerned with a local learning method of a multi-objective behavior coordination for a mobile robot. The multiobjective behavior coordination plays a role in integrating outputs of basic behavioral modules. A behavioral weight is assigned to each behavioral module represented by fuzzy rules, production rules, and so on. By updating these behavioral weights, the mobile robot can take a multi-objective situated action. However, the coordination rule is designed suitably static environments and the mobile robot must learn or update coordination rule in dynamic environments with moving obstacles. Therefore, we propose a local episode-based learning which is a learning method using self-reference of the relationship between previous perception and action in short-term memory.

This paper discusses fault tolerance in perception-based robotics from the viewpoint of ecological psychology. A prediction-based sensory network using neural networks is proposed for detecting a fault in sensing systems. Furthermore, a transformation matrix is applied for extracting perceptual information from the sensory inputs that might include fault inputs owing to breakdown. We apply the proposed method to a mobile robot. Computer simulations show the proposed method can detect the fault of sensors and can extract perceptual information used for decision making.

This paper proposes a learning method of a mobile robot with structured intelligence in a changing environment. Modular neural networks are applied for action control based on perceiving-acting cycle of ecological psychology. The robot extracts action rules from the behavior knowledge described by fuzzy rules. Next, we conduct several experiments using a mobile robot. The experimental results show the robot can learn actions based on the perceiving-acting cycle.

Proposes a controlling method of a mobile robot with structured intelligence. Modular neural networks are applied for action control based on a perceiving-acting cycle of ecological psychology. In the proposed method, the perceptual system and action system restrict each other. Next, we conduct several experiments using a mobile robot we developed. The experimental results show the robot can learn actions based on the perceiving-acting cycle. Finally, we discuss the unit of the action using modular neural networks for robotic systems.

The paper deals with behavioral evolution of multiple robots in a quasi-ecosystem. An ecosystem model composed of insects and plants, which are in a relationship of parasitism, is simulated in discrete cell space. In this ecosystem, the plants are easy to eliminate as the population size of the insects increases. Consequently, it is necessary to numerical balance plants and insects in the quasi-ecosystem. Therefore, multiple robots are introduced to remove some insects from the quasi-ecosystem. However, if the robots eliminate all the insects, viruses will eliminate the plants owing to diseases. In this ecosystem with complicated relationships, the robots should acquire strategies to maintain plants. We use simple if-then rules and apply genetics-based machine learning for acquiring a strategy for removing insects. Furthermore, we show several simulation results of behavior learning of multiple robots.

This paper deals with a prediction system based on genetic programming and fuzzy inference system. In real problems with many parameters, the prediction performance depends on the feature extraction and selection. These processes are performed using methods of multivariate statistical analysis by human operators. However, we should automatically perform feature extraction and selection from many measured data. This paper applies genetic programming for the feature extraction and selection, and further use fuzzy inference for the building energy load prediction. The functions generated by GP translate the measured data into the meaningful information that is used as input data to the fuzzy inference system. The simulation results show that the proposed method can extract meaningful information from the measured data and can predict the building energy load of the next day.

Deals with a pet robot with an emotional model. The robot requires several capabilities, such as perceiving, acting, communicating and surviving. Furthermore, it should learn various behaviors through interaction with its owner. This paper focuses on teaching a pet robot tricks or to dance. Basically, the owner can teach these tricks by simple communication based on trial and error. The robot performs the tricks by using a fuzzy controller, and further acquires tricks by a delta rule for online learning and a genetic algorithm for off-line learning. We use "Rag Warrior" as our pet robot. Experimental results show that this robot performs tricks through interaction with its owner.

This paper deals with a multi-objective behavior coordinate for a mobile robot using fuzzy control and neural network. A task given to a mobile robot includes various objectives such as collision avoiding, target tracing, and wall following. We apply fuzzy control for describing each behavior of the robot. However, a behavior might share some fuzzy rules with other behaviors. Therefore, this paper proposes a reconfiguring method for a set of fuzzy rules. The combination of fuzzy rules is updated dynamically by a neural network according to the perceptual information. Furthermore, this paper describes a learning method of the neural network and fuzzy rules based on error functions. Simulation results show that the robot can take multi-objective behavior by the proposed method.

The paper deals with a genetic algorithm for acquiring adaptive behaviors of a fuzzy based mobile robot. If its environmental state is stable or fixed, the behaviors of the robot can be optimized by conventional genetic algorithms. Otherwise, the behavior should be tuned by adaptation and learning according to the change of its environment. However, it is difficult for the robot to maintain behaviors suitable to various environmental states in the dynamic environment. Therefore, the paper proposes a genetic algorithm based on the perceived information about the dynamic environment, which is called a perception based genetic algorithm. We apply the proposed method to collision avoidance behaviors of the mobile robot in a dynamic environment. Furthermore, we conduct several computer simulations. Simulation results show that the proposed method can maintain various behaviors according to environmental changes.

This paper deals with an ecological model on planar gird of a genetic algorithm based on virus theory of evolution (E-VE-GA). In the E-VE-GA, each individual is placed on a planar grid and genetic operators are performed between neighborhoods. The E-

This paper deals with a fuzzy-based intelligent robotic system that requires various capabilities normally associated with intelligence. It acquires skills and knowledge through interaction with a dynamic environment. Subsumption architectures, behavior-based artificial intelligence, and behavioral engineering for robotic systems have been discussed as new technologies for intelligent robotic systems. This paper proposes a robotic system with "structured intelligence". We focus on a mobile robotic system with a fuzzy controller and propose a sensory network that allows the robot to perceive its environment. An evolutionary approach improves the robot's performance. Furthermore, we discuss the effectiveness of the proposed method through computer simulations of collision avoidance and path-planning problems.

Intelligent systems are required in knowledge engineering, computer science, mechatronics and robotics. This paper discusses the machine (system) intelligence from the viewpoints of adaptation, learning and evolution of living things. Next, this paper introduces computational intelligence including neural network, fuzzy system, and genetic algorithm. Finally, this paper discusses structured intelligence integrating intuitive and logical inferences, planning, and learning.

Deals with fuzzy scheduling problems including a path planning problem. We have proposed self-organizing manufacturing systems which are composed of autonomous modules. Each module decides output through the interaction with other modules, but the module does not share complete information concerning other modules. The manufacturing procedure can be divided into the sequence of three modules: tool locating module, scheduling module, and path planning module. We first solve the scheduling problem where the processing time is fuzzy. We solve the fuzzy scheduling problem by a genetic algorithm. After preplanning, the path planning module transports materials and products. Based on the processing time, the schedule module updates fuzzy processing time. We discuss the effectiveness of the proposed method through the computer simulation results.

This paper deals with an application of a virus-evolutionary genetic algorithm (VEGA) to hierarchical trajectory planning of a redundant manipulator. The hierarchical trajectory planning is composed of a trajectory generator and position generator. The position generator generates collision-free intermediate positions of the redundant manipulator. The trajectory generator generates a collision-free trajectory based on some intermediate positions sent from the position generator. To generate a collision-free trajectory of the redundant manipulator, the VEGA is applied to the hierarchical trajectory planning only based on forward kinematics. The VEGA realizes horizontal propagation and vertical inheritance of genetic information in a population of candidate solutions. The main operator of the VEGA is a reverse transcription operator, which plays the roles of a crossover and selection simultaneously. In this paper, self-adaptive mutation is applied to the VEGA for local search of trajectory planning to obtain higher performance and the quick solution. Simulation results of the hierarchical trajectory planning show that the VEGA can generate a collision-free trajectory.