Unsupervised domain adaptation (UDA) is a machine learning approach designed to minimize reliance on labeled data by aligning features between a labeled source domain and an unlabeled target domain, thereby reducing feature discrepancies, which is efficient for multivariate time series (MTS) prediction. However, most MTS UDA methods focus solely on aligning intra-series temporal features, overlooking the valuable information in inter-series dependencies. Research has highlighted that analyzing decomposed frequency dependencies in time series can reveal significant trends, noise patterns, and intricate temporal details. To address these unexplored frequency dependencies, we introduce the Frequency Graph Discovery Module (FGD), which uncovers and aligns shared frequency information and correlations across domains. Additionally, we propose a Frequency-Contextual Contrastive Learning (FCCL) framework to better capture and align frequency-contextual representations in multivariate time series, ensuring the extraction of label-invariant information for prediction. Furthermore, considering existing models overlooking the valuable and abundant information outside source and target dataset, we enhance the MTS UDA prediction model with a Language-guided Adversary Alignment (LAA) module, which leverages the advancement and capabilities of Large Language Models (LLMs) to get text-encoded labeled embeddings and align the classification features, thereby improving prediction accuracy. Our model achieves state-of-the-art results on three public multivariate time-series datasets for unsupervised domain adaptation, as demonstrated by empirical evidence.

Visual anomaly detection is an essential component in modern industrial manufacturing. Existing studies using notions of pairwise similarity distance between a test feature and nominal features have achieved great breakthroughs. However, the absolute similarity distance lacks certain generalizations, making it challenging to extend the comparison beyond the available samples. This limitation could potentially hamper anomaly detection performance in scenarios with limited samples. This article presents a novel sparse feature representation anomaly detection (SFRAD) framework, which formulates the anomaly detection as a sparse feature representation problem; and notably proposes an anomaly score by orthogonal matching pursuit (ASOMP) as a novel detection metric. Specifically, SFRAD calculates the Gaussian kernel distance between the test feature and its sparse representation in the nominal feature space for anomaly detection. Here, the orthogonal matching pursuit (OMP) algorithm is adopted to achieve the sparse feature representation. Moreover, to construct a low-redundancy memory bank storing the basis features for sparse representation, a novel basis feature sampling (BFS) algorithm is proposed by considering both the maximum coverage and the optimum feature representation simultaneously. As a result, SFRAD incorporates both the advantages of absolute similarity and linear representation; and this enhances the generalization in low-shot scenarios. Extensive experiments on the MVTec anomaly detection (MVTec AD), Kolektor surface-defect dataset (KolektorSDD), Kolektor surface-defect dataset 2 (KolektorSDD2), MVTec logical constraints anomaly detection (MVTec LOCO AD), Visual anomaly (VISA), Modified national institute of standards and technology (MNIST), and CIFAR-10 datasets demonstrate that our proposed SFRAD outperforms the previous methods and achieves state-of-the-art unsupervised anomaly detection performance. Notably, significantly improved outcomes and results have also been achieved on low-shot anomaly detection. Code is available at https://github.com/fanghuisky/SFRAD.

This paper introduces a gradient-based approach for reducing the dimensionality of acoustic features, tailored for supervised deep learning models used in speech emotion recognition (SER). This method allows us to pinpoint the crucial acoustic features that the network heavily relies on, enabling us to simplify and retrain the network accordingly. It significantly boosts testing speed, making real-time SER systems suitable for embedded systems with resource constraints in speech processing units. The proposed method is evaluated on four convolutional neural network (CNN)-based deep learning models, and one of the best results demonstrates a 56.96% reduction in test time, albeit with a minor 3.81% drop in test accuracy. The method is compared with three mainstream dimensionality reduction techniques across various dimensions, consistently outperforming them in most scenarios. A Python implementation of the method is available at https://github.com/hxwangnus/Grad-based-Dim-Red-for-SER.git.

Data-driven models for predicting the Remaining Useful Lifetime (RUL) have gained popularity due to their efficiency to enhance industrial security and reduce economic losses. Recently, there has been a notable rise in the research of transformer-based models for RUL prediction. While transformer-based models have shown significant improvements over previous LSTM-based and CNN-based models, we have raised concerns regarding high computational complexity, in-effective training with low data, no sensitivity to the order of the time series, and permutation invariant on its application to RUL prediction. The persistent issue of data scarcity and the importance of capturing the temporal relations in RUL prediction further question the suitability of transformer-based models. Considering these, We propose a simple non-transformer model, Compressed Temporal and Compressed Spatial (CTCS) Attention, which is efficient and lightweight, to capture both temporal and spatial information with the incorporation of pre- and post-positional encodings. Additionally, we introduce an Augmentation Fusion Module (AFM) to enhance the comprehension ability of the invariant characteristics of the data. The proposed methodology is evaluated on the NASA Commercial Modular Aero-Propulsion System Simulation (C-MAPSS) dataset and comprehensive experiments show that our proposed method not only surpasses other methods but outperforms the transformer-based model while requiring significantly fewer Floating-point operations (FLOPs), up to 32 times less.

Warehouse management is a critical element in the supply chain. The rise of e-commerce demands ‘any-mix any-volume’ customized and personalized orders. Due to the variety of products in ‘any-mix any-volume’ scenario the warehouse operations have become complex. The advent of multi-channel commerce has reduced the minimum order quantity (MOQ), claiming frequent changes. The ‘any-mix any-volume’ operations require flexible automation for product storage, retrieval, packing and shipping. Seamless operations of multiple devices are essential for automating flexible warehouses. For effective flexible operations, the Warehouse Management System (WMS) requires controlling the automation devices and associated auxiliary systems. Heterogenous automation devices with distinct communication protocols, data formats and connectivity are challenges for WMS centered control. An interoperable WMS can enable seamless communication and data exchange among devices. In this work, the interoperability of the WMS and various aspects of interoperability of engineering systems are presented. Different WMS system designs are considered and evaluated for the interoperability requirements. A framework for evaluating the interoperability of a WMS is proposed which is validated through experimental trials at an industry grade warehouse.

Incremental few-shot learning is highly expected for practical robotics applications. On one hand, robot is desired to learn new tasks quickly and flexibly using only few annotated training samples; on the other hand, such new additional tasks should be learned in a continuous and incremental manner without forgetting the previous learned knowledge dramatically. In this work, we propose a novel Class-Incremental Few- Shot Object Detection (CI-FSOD) framework that enables deep object detection network to perform effective continual learning from just few-shot samples without re-accessing the previous training data. We achieve this by equipping the widely-used Faster-RCNN detector with three elegant components. Firstly, to best preserve performance on the pre-trained base classes, we propose a novel Dual-Embedding-Space (DES) architecture which decouples the representation learning of base and novel categories into different spaces. Secondly, to mitigate the catastrophic forgetting on the accumulated novel classes, we propose a Sequential Model Fusion (SMF) method, which is able to achieve long-term memory without additional storage cost. Thirdly, to promote inter-task class separation in feature space, we propose a novel regularization technique that extends the classification boundary further away from the previous classes to avoid misclassification. Overall, our framework is simple yet effective and outperforms the previous SOTA with a significant margin of 2.4 points in AP performance.

Prediction of Remaining Useful Lifetime (RUL) in the modern manufacturing and automation workplace for machines and tools is essential in Industry 4.0. This is clearly evident as continuous tool wear, or worse, sudden machine breakdown, will lead to various manufacturing failures which would clearly cause economic loss. With the availability of deep learning approaches, the great potential and prospect of utilizing these for RUL prediction have resulted in several models which are designed (for RUL prediction) driven by operation data of manufacturing machines. Current efforts in these which are based on fully-supervised models heavily rely on the data labeled with their RULs. However, in these cases, the required RUL prediction data (i.e. the annotated and labeled data from faulty and/or degraded machines) can only be obtained after the machine break-down occurs. The scarcity of broken machines in the modern manufacturing and automation workplace in real- world situations increases the difficulty of getting such sufficient annotated and labeled data. In contrast, the data from healthy machines (and which are currently in operation) is much easier to be collected. Noting this challenge and the potential for improved effectiveness and applicability, we thus propose (and also fully develop) a method based on the idea of masked autoencoders which will utilize unlabeled data to do self-supervision. In thus the work here, a noteworthy masked self-supervised learning approach is developed and utilized; and this is designed to seek to build a deep learning model for RUL prediction by utilizing unlabeled data. The experiments to verify the effectiveness of this development are implemented on the C-MAPSS datasets (which is collected from the data from the NASA turbofan engine). The results rather clearly show that our development and approach here performs better, in both accuracy and effectiveness, for RUL prediction when compared with approaches utilizing a fully- supervised model.

3D part segmentation is an essential step in advanced CAM/CAD workflow. Precise 3D segmentation contributes to lower defective rate of work-pieces produced by the manufacturing equipment (such as computer controlled CNCs), thereby improving work efficiency and attaining the attendant economic benefits. A large class of existing works on 3D model segmentation are mostly based on fully-supervised learning, which trains the AI models with large, annotated datasets. However, the disadvantage is that the resulting models from the fully-supervised learning methodology are highly reliant on the completeness (or otherwise) of the available dataset, and its generalization ability is relatively poor to new unknown/unseen segmentation types (i.e., further additional so-called novel classes). In this work, we propose and develop a noteworthy few-shot learning-based approach for effective part segmentation in CAM/CAD; and this is designed to significantly enhance its generalization ability, and our development also aims to flexibly adapt to new segmentation tasks by using only relatively rather few samples. As a result, it not only reduces the requirements for the usually unattainable and exhaustive completeness of supervision datasets, but also improves the flexibility for real-world applications. In the development, drawing inspiration from the pertinent and interesting work described in the open literature as the attMPTI network, we propose and develop a multi-prototype approach (with self-attention mechanics) for few-shot point cloud part segmentation. As further improvement and innovation, we additionally adopt the transform net and the center loss block in the network. These characteristics serve to improve the comprehension for 3D features of the various possible instances of the whole work-piece and ensure the close distribution of the same class in feature space. Moreover, our approach stores data in the point cloud format that reduces space consumption, and which also makes the various procedures involved have significantly easier read and edit access (thus improving efficiency and effectiveness and lowering costs).

We aim to tackle the challenging Few-Shot Object Detection (FSOD), where data-scarce categories are presented during the model learning. The failure modes of FasterRCNN in FSOD are investigated, and we find that the performance degradation is mainly due to the classification incapability (false positives) caused by category confusion, which motivates us to address FSOD from a novel aspect of classification refinement. Specifically, we address the intrinsic limitation from the aspects of both architectural enhancement and hard-example mining. We introduce a novel few-shot classification refinement mechanism where a decoupled Few-Shot Classification Network (FSCN) is employed to improve the final classification of a base detector. Moreover, we especially probe a commonly-overlooked but destructive issue of FSOD, i.e., the presence of distractor samples due to the incomplete annotations where images from the base set may contain novel-class objects but remain unlabelled. Retreatment solutions are developed to eliminate the incurred false positives. For FSCN training, the distractor is formulated as a semi-supervised problem, where a distractor utilization loss is proposed to make proper use of it for boosting the data-scarce classes, while a confidence-guided dataset pruning (CGDP) technique is developed to facilitate the few-shot adaptation of base detector. Experiments demonstrate that our proposed framework achieves state-of-the-art FSOD performance on public datasets, e.g., Pascal VOC and MS-COCO.

In high volume production processes such as injection molding, inline inspection of parts is generally not feasible due to additional sensor requirements that incur costs. Production facilities often inspect for defective parts at the lot-level after multiple production runs and not for each run. Individual lots are accepted or rejected based on the manufacturer’s acceptable quality levels determined by the number of faulty parts in a lot. In this paper, a Lot-level Convolutional Neural Network (L-CNN) is proposed that implements two variants to improve quality estimation at lot-level using run-level sensor data. Layer-wise L-CNN uses a separation of layers within the model architecture to extract relevant features for each run. Custom loss L-CNN utilizes a standard CNN architecture with a custom loss function to handle the data’s multi-input-single-output structure. The model is evaluated using data from injection molding and wafer testing applications in the semiconductor industry. L-CNN achieves better F1and G-Mean scores compared with existing benchmarks. Even though trained with partial information (i.e., lot-level quality), custom loss L-CNN provides both lot-level and run-level quality predictions.

Dynamic multiobjective optimization requires the robust tracking of varying Pareto-optimal solutions (POS) in a changing environment. When a change is detected in the environment, prediction mechanisms estimate the POS by utilizing information from previous populations to accelerate search toward the true POS. To achieve a robust prediction of POS, a mixture-of-experts-based ensemble framework is proposed. Unlike existing approaches, the framework utilizes multiple prediction mechanisms to improve the overall prediction. A gating network is applied to manage switching among the various predictors based on performance of the predictors at different time intervals of the optimization process. The efficacy of the proposed framework is validated through experimental studies based on 13 dynamic multiobjective benchmark optimization problems. The simulation results show that the proposed framework improves the dynamic optimization performance significantly, particularly for: 1) problems with distinct dynamic POS in decision space over time and 2) problems with highly nonlinear decision variable linkages.

Accurate models for estimation of remaining useful life (RUL) is a key tool to improve productivity and reliability in modern industries. Recent advancements in machine learning and easier availability of data are driving development of data driven prognostic frameworks for estimation of RUL. Direct adoption of learning algorithms is challenging due to inherent characteristics of the domain like non-causal definition of required target, heterogeneity among the units, and temporal dependencies. This paper proposes a prognostic framework that attempts to overcome these using a modified support vector regression (SVR) as its base. SVR is modified with temporal features and an adaptive penalty term that gives higher importance to targets closer to the end-of-life of units. An ensemble of such models is learned on groups of units having similar trajectories. The framework is evaluated on benchmark data set to estimate the RUL of aircraft engines.

Factories are increasingly pushing towards automation and data-centric approaches under the current Industry 4.0 standards. Early-stage product quality estimation is identified as one of the solutions that can significantly reduce manufacturing cost and wastage. However, quality estimation is a classification problem that is inherently challenging for traditional data-driven algorithms due to its imbalanced nature. In this paper, a framework is proposed, that combines the feature extraction capabilities of convolutional neural networks and the domain knowledge characteristics of fuzzy systems. The proposed method addresses data imbalance using an ensemble of class-biased individuals, that learn features using a class-weighted loss function. Experiments were conducted using a benchmark dataset and production data acquired from the semiconductor industry. Improvements were noted for G-Mean and ROC-AVC values when compared to existing algorithms.

Autonomous robots such are unmanned aerial and unmanned ground vehicles are increasingly utilized in patrolling, surveillance, search and rescue and in missions that are hazardous for humans. Path-planning, path-generation and following a planned path successfully are fundamental requirements for autonomous operation of unmanned robots. Though the operational principle of aerial and ground robots are different, the algorithms for path-planning and path-following can be generalized. Vector Directed Path Generation and Tracking (VDPGT) proposed in this work is a platform independent path-generation and path-following algorithm. VDPGT is designed to dynamically adapt the shortest path to a destination. Simulation studies carried out on two ground robots (Turtlebot and Clearpath Husky), two aerial robots (AR Drone and Hector-quadrotor) and realtime experiments on Turtlebot and AR Drone demonstrate the platform independent nature of VDPGT.

This work demonstrates the application of Unmanned Aerial Vehicles (UAVs) in a robotic Urban Search and Rescue (USAR) mission. An intuitive and natural control methodology that can be universally applied to different types of unmanned rotary wing aerial vehicles is developed. The control system encompasses a Windows computer system and a 3-Dimensional depth-sensing camera, in particular the Microsoft Kinect. Human hand gestures are captured, filtered and translated into flight manoeuvres for the UAV. The filtering involves the application of Fuzzy Logic Controllers, which incorporate the use of several different input variables for the control of either the velocity or position of the UAV. The developed system is implemented and tested on an AR Drone.

Evolutionary algorithms are effective in solving static multiobjective optimization problems resulting in the emergence of a number of state-of-the-art multiobjective evolutionary algorithms (MOEAs). Nevertheless, the interest in applying them to solve dynamic multiobjective optimization problems has only been tepid. Benchmark problems, appropriate performance metrics, as well as efficient algorithms are required to further the research in this field. One or more objectives may change with time in dynamic optimization problems. The optimization algorithm must be able to track the moving optima efficiently. A prediction model can learn the patterns from past experience and predict future changes. In this paper, a new dynamic MOEA using Kalman filter (KF) predictions in decision space is proposed to solve the aforementioned problems. The predictions help to guide the search toward the changed optima, thereby accelerating convergence. A scoring scheme is devised to hybridize the KF prediction with a random reinitialization method. Experimental results and performance comparisons with other state-of-the-art algorithms demonstrate that the proposed algorithm is capable of significantly improving the dynamic optimization performance.

This paper explores how stimulus-dependent habituation can be integrated to modulate a cognitively inspired bottom up learning model of semantic representation. We first describe the equations governing habituation and associative habituation and describe how to integrate it to a bottom up hierarchical habituation model. Habituation is used to modulate the weights between the parent and children in the hierarchy. We investigate the behavior of habituation in the hierarchy, the properties of habituation and recovery in the model and show that while habituation has no detrimental effect on clustering, it reduces the number of neurons required for representation and therefore propose it as an important component of cognitively inspired memory systems.

Our paper explores the interaction of persistent firing axonal and presynaptic processes in the generation of short term memory for habituation. We first propose a model of a sensory neuron whose axon is able to switch between passive conduction and persistent firing states, thereby triggering short term retention to the stimulus. Then we propose a model of a habituating synapse and explore all nine of the behavioral characteristics of short term habituation in a two neuron circuit. We couple the persistent firing neuron to the habituation synapse and investigate the behavior of short term retention of habituating response. Simulations show that, depending on the amount of synaptic resources, persistent firing either results in continued habituation or maintains the response, both leading to longer recovery times. The effectiveness of the model as an element in a bio-inspired memory system is discussed.

An elastic graph matching algorithm using biologically inspired features is proposed for the recognition of hand postures. Each node in the graph is labeled using an image feature extracted using the computational model of the ventral stream of visual cortex. The graph nodes are assigned to geometrically significant positions in the hand image, and, the model graphs are created. Bunch graph method is used for modeling the variability in hand posture appearance. Recognition of a hand posture is done by the elastic graph matching between the model graphs and the input image. A radial basis function is used as the similarity function for the matching process. The proposed algorithm is tested on a 10 class hand posture database which consists of 478 grey scale images with light and dark backgrounds. The algorithm provided better recognition accuracy (96.35%) compared to the reported results (93.77%) in the literature.

In this paper, landing stability of jumping gaits is studied for a four-link planar biped model. Rotation of the foot during the landing phase leads to underactuation due to the passive degree of freedom at the toe, which results in nontrivial zero dynamics (ZD). Compliance between the foot and ground is modeled as a spring-damper system. Rotation of the foot along with the compliance model introduces switching in the ZD. The stability conditions for the ldquoswitching ZD rdquo and closed-loop dynamics (CLD) are established. ldquoCritical potential index rdquo and ldquocritical kinetic indexrdquo are introduced as measures of the stability of the CLD of the biped during landing. Landing stability is achieved by utilizing the stability conditions. Stable jumping motion is experimentally realized on a biped robot.

Deoxyribonucleic acid (DNA) coding methods determine the meaning of a certain character in individual chromosomes by the characters surrounding it. The meaning of each character is context dependent, not position dependent. Although position-dependent coding is most commonly used in genetic algorithms (GAs), a context-dependent coding formation is in fact more closer to the natural DNA chromosome. With the context dependency, the DNA coding methods allow intron parts, redundancy, and variable string length in encoded strings while remaining compatible with the standard genetic operations. This paper tries to explicitly explore the influence of those special features of the DNA coding scheme. Two fundamental DNA coding methods (with and without the use of introns) are constructed and compared with the integer coding method, which lacks the features of interest. The performance of the proposed DNA coding methods is analyzed through the robot soccer role assignment problem. The context-dependent coding exhibits the advantages in handling the negative effect of epistasis. The redundancy and intron parts are helpful in preventing useful schemata from disruption and in increasing the population diversity. The variable length of the individual string enables GAs to evolve both the size and the structure of the fuzzy rule base.

In this paper, we present an approach to motor skill learning based on internal models. By pursuing the temporal and spatial scalability of internal models, we first investigate the possibility of generating similar movement patterns directly via the same internal model with the minimum changes in the internal model parameters, and avoid the reinforcement learning. Next, we consider more complex movements for which different internal models are needed. Based on the task decomposition, all movements can be classified into the sequential and parallel DMPs. The former requires a number of IMs to work sequentially so that a sophisticated motor behavior can be performed. The latter also requires a number of IMs to work in parallel to generate the needed movement patterns. To mimic the human limb behavior, a two-link robot arm is used as the first prototype to perform the motor learning process of letter writing. A FUJITSU HOAP-1 humanoid robot is used as the second prototype and the upper limb movement is conducted in real-time, which further validates the effectiveness of multiple internal model approach for motor learning.

An extensive fuzzy behavior-based architecture is proposed for the control of mobile robots in a multiagent environment. The behavior-based architecture decomposes the complex multirobotic system into smaller modules of roles, behaviors and actions. Fuzzy logic is used to implement individual behaviors, to coordinate the various behaviors, to select roles for each robot and, for robot perception, decision-making, and speed control. The architecture is implemented on a team of three soccer robots performing different roles interchangeably. The robot behaviors and roles are designed to be complementary to each other, so that a coherent team of robots exhibiting good collective behavior is obtained.

The advantages and disadvantages of evolving neural control systems for mobile robots using genetic algorithms are investigated. The Khepera robot is trained using the evolutionary neural networks (ENN) algorithm for the task of obstacle avoidance. The feasibility of using Q-learning for robot learning is also studied. It is found that Q-learning can be successfully used to train a robot and is more promising than the ENN algorithm in this case. The Webots simulation software has been used to carry out all the experiments.

A comparison of mobile robot navigation using evolutionary neural networks and the pain based algorithm is discussed in this paper. The controllers are designed based on evolutionary neural networks and the pain-based algorithms. The performance of the controllers are verified with the Khepera robot.

A DNA coded genetic-algorithm (GA) is proposed to optimize the rule-base of a fuzzy logic controller (FLC). The controller is designed for a vehicle-active suspension system to improve the driving comfort. The DNA codes GA constructed optimal decision-making rules for the fuzzy logic controller. Simulation results demonstrate the effectiveness of the algorithm.

A new methodology named Evolutionary Artificial Potential Field (EAPF) is proposed for real-time robot path planning. The artificial potential field method is combined with genetic algorithms, to derive optimal potential field functions. The proposed EAPF approach is capable of navigating robot(s) situated among moving obstacles. Potential field functions for obstacles and goal points are also defined. The potential field functions for obstacles contain tunable parameters. The multi-objective evolutionary algorithm (MOEA) is utilized to identify the optimal potential field functions. Fitness functions such as goal-factor, obstacle-factor, smoothness-factor and minimum-pathlength-factor are developed for the MOEA selection criteria. An algorithm named escape-force is introduced to avoid the local minima associated with EAPF. Moving obstacles and moving goal positions were considered to test the robust performance of the proposed methodology. Simulation results show that the proposed methodology is efficient and robust for robot path planning with non-stationary goals and obstacles.

A real time vector field based path planning for an attack mode robot and a Petri net state diagram approach for the robot's role selection for robot soccer games are proposed in the paper. A robot soccer game has a dynamic environment as different robots intentionally affect the environment in unpredictable ways in a competitive setting. A soccer-playing robot has to take appropriate action according to its surroundings. The efficiency and applicability of the proposed controllers were demonstrated through real robot soccer games in MiroSot'97, held at KAIST Korea, June 1-5, 1997.