Although prior research has examined dyadic relationships between children and robots, empirical evidence is lacking on how robots can support childcare by fostering relationships involving both children and mothers. This study advances human–robot interaction research by comparing families that lived with a robot for 7 days (robot-present group) with those that did not (robot-absent group) to assess whether robots reduce stress during mother–child separation. By integrating physiological and behavioral measures, we examined stress reduction in 44 mother–child pairs ( n = 18 and n = 26, respectively). We measured the cardiac acceleration capacity, a specific indicator of sympathetic stress response, along with children’s heart rate, movement distance, and duration of physical contact with the robot during a separation–reunion procedure. The results indicated that (1) mothers in the robot-present group exhibited lower sympathetic activation during separation than those in the robot-absent group; (2) children in the robot-present group displayed a decreased heart rate despite increased movement during separation; and (3) children significantly increased their physical interaction with the robot during separation. These findings suggest a positive feedback loop: as children utilized the robot as a stress buffer, mothers gained visual reassurance of their children’s adaptive responses, further reducing maternal stress. Theoretically, this study extends attachment theory to human–robot interactions by demonstrating that a robot can serve as a safe haven , without replacing the full attachment bond established with human caregivers. Practically, this study introduces a novel parenting support approach wherein robots offer direct support to children while simultaneously providing social support to mothers.

This study proposes a caregiver task assistance system using XR devices based on task recognition with 3D human pose estimation in an Intelligent Space. The system integrates spatial memory for organizing digital data, a 3D model for accurate caregiver posture recognition, and an AR device for real-time task guidance. Experiments with simulated care tasks demonstrated the system's effectiveness in reducing caregivers' anxiety about forgetting subtasks and improving overall task performance. Future work will focus on verifying the system's usefulness for elderly caregivers.

Parent-Child Interaction Therapy (PCIT) is a well established method used by psychologists to improve and to maintain a healthy and good parent-child relationship. PCIT requires trained and certified therapists, and the therapy sessions are conducted in a special room, but participants can also practice and later perform sessions in their homes with our without therapist supervision. In order to facilitate the therapy for the parents, here we present a possible approach to introduce social robotics into the PCIT therapy. We developed a behavior model to control a mobile robot and interact with both the parent and the child during a therapy session with the goal of stimulating interaction and communication between the parent and child. The behavior of this social robot depends on the contents of the conversation between the participants, which is analyzed in real-time. For speech analysis we have developed a framework for integrating various already available components to achieve real-time Speech-to-Text transcription, and also developed our own Natural Language Understanding (NLU) engine. The evaluation of the conversation is performed by this NLU engine, which was inspired by the Dyadic Parent-Child Interaction Coding System (DPICS) used in the PCIT therapy. Depending on the evaluation results, our mobile robot can exhibit different behaviors, which were inspired by EthoRobotics, such as approaching, getting attention, playing fetch with a toy. Besides constructing a functional real-time human conversation analysis system and a behavior model, we also performed preliminary HRI experiments with a real mobile robot controlled by the proposed model in real-time.

This paper proposes an approach to assist the visually impaired in walking independently by presenting people flow and sidewalk edge information. To achieve this, a robust estimation of the direction of people flow, and a method for detecting the edges of sidewalks are presented. Image learning for sidewalk estimation is performed using binary semantic segmentation that classifies sidewalks and other areas to see sidewalks regardless of pedestrians and shadows on sidewalks. The effectiveness of the methods was evaluated. Then, preliminary user experiments will investigate the effect of the walking assistance information presented by this approach on the decision of walking behavior.

Inspired by a common behaviour pattern from the dog-human relationship, we present a behaviour model combined with a method for controlling a real mobile robot, effectively realizing physical Human-Object-Robot Interaction. Translating social behaviours from the dog-human relationship into social robots has been already found to be a viable way. Playing, a cardinal behaviour element in the dog-human relationship, turned out to be one of the highest priority factors in forming a successful and lasting relationship between dogs and humans. Accordingly, we modeled the playing behavior when a human repeatedly initiates playing with a dog by picking up and throwing an object (toy), expecting the dog to go for the object and bring it back to the person throwing it, commonly known as playing fetch. We also developed a robot control system able to execute the commands from the behaviour model, and for gathering the poses of the participants in the real physical space we have used the What-You-See-Is-What-You-Get Indoor Localization (WIL) system. The integration of these components allowed us to create a social robotics system able to offer real physical human-object-robot interaction fast enough to be enjoyable for human participants, as a tool to be used in future HRI experiments, which was our main goal. The proposed behaviour model is intended to be embedded into social robots with the purpose of applying it in e.g. therapy, child-rearing, child-robot interaction scenarios, human behaviour assessment, recreational or entertainment scenarios.

In recent years, 15% of the working population have suffered from mental disorders. Such conditions cause low productivity, absence of diseases, and unemployment. In addition, the treatment of mental disorders takes a long time, so routine precautions are essential. One of the ways to prevent mental disorders is to achieve healing through interaction with animals. However, this method has not been widely accepted because of the difficulties in handling animals. To solve these problem, we introduced a virtual pet through a virtual reality technology for three-dimensional stereoscopic visualization as an alternative to animals. We named the system that prevents mental disorders by interacting with the virtual pet as virtual pet-assisted activity (VAA) and verified that this system gained short-term psychological and physiological benefits. However, it was unclear what causes the system to have positive effects on users. This study investigated whether space factors affect the psychological and physiological aspects of the user and whether the scenery factor affected the psychological and physiological aspects of the user and its interactions with the virtual pet. The experiment was divided into two experiments for the space factor and scenery factor. As a result, it was not affected by the space factor but was affected by the scenery factor. The findings reveal that the open scenery is more effective than the closed scenery in the VAA system.

According to a 2018 World Health Organization investigation, worldwide, there are around 253 million visually impaired persons. Technological support is required to help such people walk safely. This paper focuses on the fact that many pedestrians are forming a flow in actual walking environments. We believe that behavior decisions based on people flow lead to safe and reliable walking of the visually impaired people. The purpose of this study is to assist a visually impaired person's recognition of surrounding dynamic environments by presenting people flow. In this paper, we propose a method of estimating the situation of pedestrians around and direction people flow from an image. In addition, we propose a method of people flows presentation using vibrotactile stimulation. After identifying the vibration perception characteristics of the presentation positions on a human body, we design and verify the vibrotactile stimulation for people flow presentation.

In recent years, ambient assisted living (AAL) has studied and a demand for services that monitors living environments to reduce the burden of nursing care has increased. Therefore, we propose an approach of nonverbal human-robot communication for AAL applications based on an attachment behavior model for a mobile robot. This can sense the situation within the environment in real time with less burden on the user. We use an etho-robot as the user interface of an AAL system for useful monitoring services. The etho-robot mimics dogs' attachment behaviors and can provide less stressful communication between humans and robots. In addition, we design behavior for providing notification about a person's status for the etho-robot. The integration of a mobile robot as a user interface of the AAL system will make the communication spatially flexible. We conducted an experiment to evaluate whether humans can correctly recognize the attachment behavior of the etho-robot when the robot performs its notification function. We discuss the usefulness of mimicking a dog's attachment behavior when the robot has such a function.

Industrial robots are introduced in not only large enterprises but also small and medium enterprises (SMEs). Professional engineers are required when introducing and maintaining industrial robots. However, sometimes it is difficult for SMEs to do this by themselves because of economical limitations. Therefore, in many cases, they need to ask the outside for the support. To assist this situation, an aiding system for remote operation of industrial robots has been studied. This system provides information of the local site to an operator who is in the remote site using auditory and visual information. From this, the operator is able to perceive the working situation of the robot remotely. However, when several kinds of auditory information are presented at same time, recognition of the information becomes difficult. In addition, if too many kinds of visual information are presented, it is also difficult for the operator to pay attention appropriately. To solve the problem, we have proposed vibrotactile interfaces to assist information presentation for aiding remote operation of an industrial robot. In this paper, we attempt to present contact information using a vibrotactile glove when contact between an end-effector of a robot and an object occurs. We proposed two types of vibration stimuli to present the contact information. We investigate how human subjects correctly perceive the contact information through the vibration-based presentations compared with a visual based presentation. Subjects? performance of the contact information perception is evaluated based on perceived time, cognitive time and correct answer rate of the contact information.

To develop long-term human-robot communication (HRC), robots' behavior must be properly perceived and interpreted by users. However, evaluations of processes involving a robot and a human are not conducted in natural situations in which humans can freely interact and observe a robot's behavior. In this study, an experiment involving both a mental model and an impression of a robot was proposed to evaluate the effect of human expectations on HRC.

A smart electric wheelchair is proposed for use as a personal mobility tool (PMT) . To reduce the load on the user, PMT is integrated with autonomous mobile robot navigation and human inputs. In addition, an obstacle avoidance function is provided. To facilitate comfortable and easy movement as well as safety and efficiency, we aim to realize a PMT with obstacle avoidance that does not cause discomfort and frustration to the user. The tool detects a human on the basis of the leg shape. The virtual impedance method is used as an obstacle avoidance method in dynamic environments. An experiment is conducted to evaluate the psychological effect of the PMT's obstacle avoidance movement when different types of obstacles and movement characteristics are applied.

Robots that live with a person for a long term need to build social relationships with the user. We focus on the human-dog relationship because a dog will be attached to the human owner. The dog and the human will then build social relations with each other. This paper proposes an approach to update the relationship between a person and a robot based on the experiences of communication with the person. The relationship between the person and robot is expressed using three types of behavioral factors: degree of attachment with the owner, degree of acceptance to a non-owner, and degree of sensitivity of anxiety. The three behavioral factors determine the behavioral characteristics of the robot. The problem with the previous model is that the parameters of the factors are given as fixed numerical values. As a result, the relationship between persons and robots never change, even though a person treats the robot with affection. More specifically, no mechanism is used in the previous model to change the robot behaviors according to the history of human-robot communication, which causes problems. For example, persons cannot appropriately interpret and estimate the relationship with the robot because the robot behaviors never change. In this paper, we propose a mechanism to update the relationship between persons and a robot by evaluating the human reactive behavior against the robot behavior. Moreover, we evaluate whether human subjects can find and interpret the change in the relationship with the robot using the proposed method.

With the trend of increasing human robot collaboration, the need for more advanced safety systems arises. To create safe robots that act proactive to dangers, a safety system based on risk analysis has been developed. This paper presents a vital part of the risk analysis, the likelihood analysis,. In this likelihood analysis, data about human activity is recorded based on its occupied space. The times of an occupation are recorded in absolute and relative time. A multi-peak probability distribution is fitted to the data using a hidden genes genetic algorithm. This algorithm enabled the system to also optimize the number of peaks in the distribution. The system is implemented and simulated for a simple assembly task. Plots from the simulations show the system's ability to predict active and inactive periods of time in a given area. This information can later be used in a risk analysis based scheduling problem.

The first part of a cone's signal transduction is investigated from an image processing perspective in order to find out what differentiates (human) vision from computer vision. We found that the activity of cone opsins— visual pigments that are activated by the impact of a photon—can be described as an approximation of a fractional integrator of order 0.1–0.2 on frequencies between 1–30 Hz. We explore how this affects the output signal and provide examples of how this can be used for noise reduction and image processing. We also present a simplified model since these processes require excessive computational power for computer vision modeling.

Recently, social problems associated with a sense of isolation among elderly adults have been highlighted. To address this problem, we propose a long-term human-robot communication program inspired by human-dog relationships. An Etho-engine was improved upon to provide robot accommodations for access within rural regions. We generalized the robot's inner states and updated rules governing decision-making based on interactions with external situations. Additionally, to develop robot preferences for a person other than the owner, a "non-owner" was defined as being initially unfamiliar, with gradual acceptance being updated according to interaction histories and relationships between the non-owner and the owner. Experiments were conducted to evaluate the communication model. We discuss the effectiveness of this modified Etho-engine based on the experimental results.

This study addressed the leading behavior of a robot. To lead a person when the person's attention is initially elsewhere, the robot's behavior should be designed such that it seeks the person's attention and seamlessly brings him or her to a target location. Therefore, we set out to design and implement a leading behavior for a robot inspired by the dogs' behavior sequence. In particular, we considered the legibility of the movements used by the robot to show a robot's destination clearly. We evaluated the autonomous robot's leading behavior through an experiment.

To sustain effective and long-term human-robot communications, we propose a robot behavior model based on dog behavior in human-dog relationships. The model includes behavioral factors that produce different behavioral characteristics. In this study, we improved the "play with the owner" behavior, which is the only mutual interaction between a person and the robot among the robot's attachment behaviors. We evaluated the attachment behavior with the improved "play with the person" behavior.

A smart electric wheelchair for use as a personal mobility tool (PMT) is proposed. To reduce the load on the user, the PMT is integrated with autonomous mobile robot navigation and human input. It is able to reach a destination autonomously based on a directional input from the user. The accuracy of the destination detection depends on the user's input. Therefore, to realize stress-free and efficient movement of the PMT, an appropriate means of destination detection based on inaccurate human input is required. We propose an approach whereby the PMT can determine its destination based not only on the user's input but also on a history of human activity in the living environment. To enable the use of a history of human activity to determine a destination, this paper proposes a map-building method to express a history of human activity on a grid map. An experiment is conducted to evaluate the usefulness of the proposed approach for destination determination based on the subjects' performance of the PMT operation and their mental workload.

To sustain effective and long-term human-robot communications, a robot behavior model based on dog behavior in human-dog relationships is proposed. The model includes behavioral factors that produce different behavioral characteristics. We conducted experiments to evaluate robot-user impressions of robot behaviors when presented with different characteristics using a simulator. Our findings showed that the robot-users had the impressions that we expected based on the factors selected. As in an earlier model, the behavioral characteristics were assigned fixed numeric values; therefore, the characteristics in that earlier model could not be changed by users. This paper, however, proposes a method that enables users to modify or “tune” the behavioral characteristics based on verbal communication.

We investigated the correspondence relationship between vibrotactile stimuli and tactile mental imagery using a vibrotactile glove proposed in a previous study as an interface for vibrotactile stimuli [1]. The mental impressions of the vibrotactile stimuli obtained by human subjects were measured by semantic differential. Analysis of variance, multiple comparisons, and factor analyses were used to extract and evaluate the distinctive imageries induced by passive touch based on the answers given to questions the subjects were asked. We also discuss the potential of the glove to produce the five fundamental sensations of fine roughness, macro roughness, hardness, warmness, and friction.

This paper presents the social behavior of a physical agent in intelligent space. Intelligent Space (iSpace) is a space in which many intelligent devices, called distributed intelligent network devices (DINDs) and including not only sensors but also actuators, are installed. Robotic devices, which provide services to users, are considered as physical agents in iSpace. For facilitating long-term communication between users and a robot, we focus on building a social relationship between a user and a robot. Toward this end, a methodology for designing social behaviors of the robot is required. Therefore, we have used an ethologically inspired approach to construct a robot behavioral model for an iSpace-based elderly monitoring support system. In this paper, the robot behavioral model is presented and implementation of robot behaviors is discussed.

To maintain long-term human-robot communication, a robot behavior model based on the behaviors of dogs in human-dog relationships has been proposed. We apply this model to a robot in a monitoring support system designed for home care. This paper presents an improvement to the model, adding behavioral factors to show different characteristics of the behaviors. We conduct experiments to evaluate users' impressions of the robot behaviors when given different characteristics using a simulator. Our findings show that the subjects had the impressions that we expected based on the factors selected.

This paper presents an exploratory behavioral model for human-robot communication based on an ethological approach. To maintain long-term human-robot communication, building social relationships between the user and robot is important. As a model of social relationships between different species, we focus on human-dog relationships. We apply dogs' behaviors observed from the stranger situation test as a base model of the robot. Exploratory behavior is one of the dogs' behaviors. In this paper, human-robot communication is considered for the purpose of home-care support. For this purpose, an exploratory behavior model for home-care applications is proposed.

We have proposed the spatial memory which is a human interface to operate iSpace agents such as mobile robots, computer displays, etc. Spatial memory enables users to store computerized information ranging from digital files and commands for robots in real space by assigning digital information to a three-dimensional position. Users can retrieve the information by pointing the locations using their own bodies directly. The accessed information will be delivered to a suitable iSpace agent who should execute it. We named information arranged into real space ldquoSpatial-Knowledge-Tagrdquo (SKT). In order to encourage people to share useful information via spatial memory, they need to retrieve pre-arranged SKTs which are arranged by somebody. In this paper, firstly we show possibility of finding SKTs through seeing other users' actions, but not using any display to show locations and content types of SKT. Then, we present two kinds of displays to show users locations and content types of SKTs. Specifically, one is a visualization using computer graphics (we call ldquographical displayrdquo), the other is a display using vibration (we call ldquovibration displayrdquo). Experiments to compare the efficiency of the methods to find SKTs are shown.

This paper describes an observation system of human-objects interactions to extract relations among humans and physical objects in Intelligent Space (iSpace). The purpose is to obtain information of physical objects which can not be described in advance, in other words, depends on usage history of individual users. It is important issue for the iSpace to recognize human activities and design services for humans. In this paper, a method for multiple objects localization in the Intelligent Space is presented. The localization is realized by using two types of information. One is acceleration data which is used to determine whether an object is being operated. Acceleration data is acquired by using sensor network system. The other is hand position information of a nearest user to one object. By combining the information, positions of objects are able to be estimated by corresponding positions of user hands. In this method, determination of objects initial position is a first problem. In order to solve it, we use correlation of acceleration data among user hands and objects. Initial position of each object is given based on user’s hand position which is scored high correlation.

This paper presents a classification of human activities based on usage history of the spatial memory system and a numerical expression of each human activity in order to estimate individual human's intended purpose of an environment. We named the intended purposes for a place "space-human activity association". Intended purposes for a place will be different depending on the purpose of activities, even if observed activities are almost the same. Therefore, estimation of space-human activity association is important for intelligent environments to design services provided for individual human according to his/her current situation. The spatial memory system enables human users to store computerized information into the real world by assigning a three-dimensional position to the information, and to retrieve the information by directly indicating the point using their own hands. In the spatial memory system, what associates computerized information with a three-dimensional position is called "spatial-knowledge-tag (SKT)". The users actively create SKTs based on their need. As a result, arranged SKTs in a specified environment correspond to activity histories of the users. Also, users' intended purposes for the environment are reflected in the arrangement. Therefore, classification of arranged SKTs leads to classification of human activities, and accordingly estimation of intended purposes for places. In this paper, we describe two approaches to classify human activities. More specifically, the first approach is a classification of arranged SKTs by using only SKTs' information. The second approach is a classification of arranged SKTs by using both SKTs' information and usage history of the spatial memory system. As a result, the classification based on usage history of the spatial memory system showed multiple activities in the same area. On the other hand, the first approach resulted in a single cluster in the same area. We, then, give numerical values to SKT's content type, and obtained numerical expression of each human activity

This paper proposes the spatial memory as an interface between human and intelligent space (iSpace). The spatial memory regards three-dimensional space as mass storage of computers, i.e. three-dimensional point is treated as an address of stored digital data such as various documents, images and commands for machines. Consequently, we can access stored digital data by indicating a point using our own arms, which we named "human indicator". In this paper, we implement a prototype of the spatial memory system which is supposed to utilize digital data regarded as externalized knowledge. The implemented system is evaluated through two types of experiments. The tendency to forget and the efficiency of task performance are investigated. By applying the t-test under the level of significance 5%, experimental results of the performance efficiency show significant statistical difference between the results with the spatial memory and without it.

This paper describes a user-friendly system to recognize the surrounding obstacles around the elderly using the hand-haptic force feedback, to help the elderly walk safely. The user is able to attain the surrounding recognition by perceiving repulsive force generated in feedback while manipulating the joystick. To build up the system, a virtual potential fields corresponding to the distance and direction to the obstacle is employed. Through the experimental results, it is found that the practice time that the user learns basic operation of the system is sufficiently short. Furthermore, the user feels the sense of security while recognizing the surrounding information through the hand-haptic force.

This paper investigates how the human feels the moving behavior of the omnidirectional mobile robot that follows the human while the robot swingingly moves around the human. To realize the swinging motion, the modified virtual impedance method is implemented on the controller for moving behavior of the robot. The psychological evaluations are examined for the impression of the moving behavior. In the evaluation, the SD (semantic differential) method and the RRV (R-R wave variance) method using the electrocardiogram (ECG) as biosignal are employed. It is revealed that the moving behavior makes a certain impression on human. Furthermore, it is confirmed that the RRV method is available in the real time evaluation.