Knowledge distillation (KD) aims to enhance the performance of lightweight student networks through the guidance of teacher models. However, the existing methods have deficiencies in two key aspects: First, these methods rely heavily on static representation alignment, failing to account for optimization sensitivity in different directions within the distillation subspace; second, they lack a fine-grained mechanism to align critical directional features. To address these issues, we propose Direction Sensitivity–based Knowledge Distillation method (DSKD), which can quantitatively measure the sensitivity of each direction to the loss function at different training stages and dynamically select the optimization direction accordingly. Meanwhile, we designed a directional sensitivities weighted distillation loss. By aligning the parameter matrices of the teacher and student models in the key directions, we can more effectively transfer knowledge and improve the distillation effect. We combined DSKD with multiple advanced distillation strategies and conducted an empirical evaluation in the GLUE benchmark and CIFAR-100. The results showed that this method could significantly improve the performance of existing distillation techniques.

The paper proposes a novel Economic Model Predictive Control (EMPC) scheme for Autonomous Surface Vehicles (ASVs) to simultaneously address path following accuracy and energy constraints under environmental disturbances. By formulating lateral deviations as energy-equivalent penalties in the cost function, our method enables explicit trade-offs between tracking precision and energy consumption. Furthermore, a motion-dependent decomposition technique is proposed to estimate terminal energy costs based on vehicle dynamics. Compared with the existing EMPC method, simulations with real-world ocean disturbance data demonstrate the controller’s energy consumption with a 0.06% energy increase while reducing cross-track errors by up to 18.61%. Field experiments conducted on an ASV equipped with an Intel N100 CPU in natural lake environments validate practical feasibility, achieving 0.22 m average cross-track error at nearly 1 m/s and 10 Hz control frequency. The proposed scheme provides a computationally tractable solution for ASVs operating under resource constraints.

The execution of long-distance load-carrying tasks across multiple terrains remains a frequent requirement. These tasks often involve heavy loads, resulting in fatigue, decreased efficiency, and potential safety risks. To address this issue, this paper proposes a wearable centaur robot with wheel-legged transformation for human load-carrying assistance. The key feature of this robotic mechanism is the independent wheel-legged transformable structure, enabling transitions between the wheeled and legged modes. The wheeled mode ensures high load-carrying efficiency, while in the legged mode, the wheels are laid flat, transforming the ankle joint into a locked support surface that provides stable gait support. This design enables efficient and stable load carriage over complex terrains, all while preserving the natural gait of the user. Next, we develop a unified control framework for human-robot collaborative locomotion across different terrains, which includes velocity control based on an admittance model for the wheeled mode, gait control using a Bézier trajectory for the legged mode, and the transition between the two modes. The preliminary experiments include wheeled-mode, legged-mode, mode transition and obstacle crossing under human-robot collaborative locomotion, validating the proposed robot’s adaptability to different terrains while assisting with human load carriage.

In cluttered environments, a human-following mobile robot must predict the motion intention of the followed human and take environmental obstacles into consideration. Consequently, it brings several challenges, such as the human’s detour direction prediction problem and the visibility maintenance problem for route planning. To overcome these problems, this paper proposes an integrated follow-ahead framework, in which the human’s detour behavior is predicted by the Leg Motion Model-based EKF (LMM-EKF) and the iterative human route search algorithm, followed by the Safe Corridor-based Model Predictive Controller (SCMPC) used to obtain the optimal control solution. Also, a new perspective about visibility is provided in this paper that, via placing multiple obstacle-free safe regions along the human’s intended direction without any complex preprocessing for the point cloud, SCMPC prevents the robot from collision and occlusion simultaneously based on the basic properties of the convex set. The validity of the proposed method is comprehensively verified through real-world experiments.

Physiological signal-based emotion recognition requires effective fusion of multi-modal physiological signals to improve recognition accuracy. In this paper, a multi-kernel embedding fusion framework (MKEFF) is proposed for multi-modal physiological signal emotion recognition. Specifically, multi-kernel learning and kernel approximation techniques are used to compute the multi-kernel embeddings of the original feature vectors of each modality independently. The embeddings are then fed in parallel to their respective representation learning layer, where the proposed sparse relation learning method is applied to all the modalities to explore the correlation and diversity among them. Finally, a distribution alignment based fusion method is proposed to align each modality in the subspace, and a weighted summation fusion is performed to obtain the fused representations. Extensive cross-subject emotion recognition experiments are conducted on three public datasets, DEAP, DECAF, and SEED-IV, to evaluate the proposed method. The experimental results demonstrate that the proposed method achieves better classification performance and interpretability than the state-of-the-art methods.

In this work, we conducted research on deformable object manipulation by robots based on demonstration-enhanced reinforcement learning (RL). We present FADERL (Fuzzy-Augmented Demonstration-Embedded Reinforcement Learning), a novel framework for robotic manipulation of deformable objects that significantly improves reinforcement learning efficiency through synergistic unification of High-Dimensional Takagi-Sugeno-Kang (HTSK) fuzzy systems, Generative Adversarial Behavior Cloning (GABC), and Conditional Policy Learning (CPL). Compared to the Rainbow-DDPG baseline, FADERL achieves 2.01× higher global average reward and reduces standard deviation to 45% while requiring fewer computational resources. To address the high cost of human demonstration collection, we introduce a Nonlinear Model Predictive Control (NMPC)-based data augmentation method that generates high-quality demonstrations at minimal cost. Simulation results demonstrate that NMPC-generated demonstrations enable FADERL to achieve performance comparable to human demonstrations. Physical experiments on fabric manipulation tasks—diagonal folding, central-axis folding, and flattening—achieve success rates of 83.3%, 80.0%, and 96.7% respectively, validating our approach’s effectiveness in real-world scenarios. Unlike computationally intensive large-model approaches, FADERL provides a lightweight, task-specific solution with efficient adaptability, making it suitable for practical robotic applications in manufacturing, medical surgery, and service robotics.

In light of global aging and prevalent stroke-related hemiplegia, this study addresses challenges in robot-assisted Sit-to-Stand (STS) movements, a daily activity prone to falls. Supernumerary Robotic Legs (SRL) serve as independent support, enhancing stability and limb movement range. Existing coordination control methods lack personalization for STS assistance, requiring solutions for human intent transmission and rapidly optimize coordination control challenges in the non-coupled human-robot system. The proposed human-SRL coordination control algorithm, grounded in personalized SRL-human coupling models, incorporates surface electromyography (sEMG) signals to design an intent-driven variable stiffness impedance control. The inclusion of incremental learning enables rapid optimization of impedance parameters, facilitating real-time adjustments in SRL assistance for adaptive coupling with users. Practical experiments involving both healthy participants and hemiparetic patients validate the algorithm’s effectiveness during STS. The results validate substantial reductions in STS time (39.54%) and muscle activity (28.01%), highlighting the efficacy of the proposed algorithm-controlled SRL support for hemiparetic individuals.

The federated learning (FL) technique can provide a promising solution for the timely training of a deep learning model with the critical requirement of privacy protection. However, the existing FL frameworks still confront challenging issues including heterogeneous data sources, edge device heterogeneity, sensitive information leakage, nonconvex loss, and communication resource constraints which place obstacles in terms of practicality. In this article, first, a federated multitask learning (FedMTL) approach is introduced to reformulate the FL model as a multiobjective optimization problem which results in federated multigradient descent algorithm (FedMGDA) with a better model personalization against data heterogeneity and Byzantine attack. Second, a new semi-asynchronous model aggregation method is developed to asynchronously aggregate small partial clients for compensating impacts of the straggler and staleness. Third, a distributed differential privacy technique is applied to enhance the privacy protection of asynchronous FedMGDA with the convergence guarantee where the convergence analysis of differentially private asynchronous federated multiple gradient descent algorithm (DP-AsynFedMGDA) is studied for both the convex and the nonconvex loss functions. Empirical examples and comparative studies are presented to illustrate the effectiveness of the proposed DP-AsynFedMGDA.

In rehabilitation applications, encouraging patients to actively participate in training is essential for effective recovery. However, personalized control design in robot-assisted therapy remains challenging due to variations in patients' motor capabilities. To address this issue, this paper proposes an assist-as-needed (AAN) control framework that integrates a hybrid fuzzy-transformer neural network (HFTN) with a fuzzy echo state network (FESN)-based variable impedance controller to ensure personalized support and active engagement. The HFTN integrates fuzzy logic with transformer architectures in parallel paths, establishing a novel neuro-fuzzy musculoskeletal (MSK) model that maps surface electromyography (sEMG) signals to joint torque through combined uncertainty and temporal modeling for enhanced real-time estimation. The variable impedance controller constructs the stiffness and damping matrices of the robotic system through the FESN and develops an adaptive update law for the FESN output weights, effectively addressing instability issues in variable stiffness control. Furthermore, driven by physiologically estimated joint torques from the HFTN, the adaption of the FESN reservoir states enables real-time modulation of stiffness and damping, facilitating transitions between human-dominated and robot-dominated modes. This realizes the AAN concept, ensuring personalized and responsive assistance. Various experiments on an upper limb rehabilitation robot were conducted to validate the effectiveness of both the neuro-fuzzy MSK model and the AAN controller in delivering optimal assistance while promoting active user participation.

Current research on grasping state analysis in soft manipulators is limited and lacks broad applicability. In this article, we introduce a novel method that leverages multimodal data from flexible sensors and Inertial Measurement Units (IMUs) to develop a comprehensive grasping state analysis system for multidegree-of-freedom (multi-DOF) pneumatic soft manipulators. A Deep Spiking High-Dimensional Fuzzy Network (DSHTFN) algorithm is specifically designed to analyze the “3S” grasping states of soft manipulators—shaking, stable, and slipping—with greater depth and precision. A novel membership function, the BernoulliArctangent (B-Atan) function, has been designed to accommodate the unique characteristics of spiking input signals and support backpropagation capabilities. Experimental results demonstrate that our proposed method achieves accuracies of 95.66% and 96.05% in opposing-finger and three-fingered soft manipulator operations, respectively. Through comparative analysis with other algorithms, the superior performance of the B-Atan membership function and the DSHTFN approach in analyzing the grasping states of soft manipulators has been validated.

Motor imagery (MI) decoding is the basis of external device control via electroencephalogram (EEG). However, the majority of studies prioritize enhancing the accuracy of decoding methods, often overlooking the magnitude and computational resource demands of deep learning models. In this study, we propose a novel lightweight Multi-Scale Feature Residual Convolutional Neural Network (MFRC-Net). MFRC-Net primarily consists of two blocks: temporal multi-scale residual convolution blocks and cross-domain dual-stream spatial convolution blocks. The former captures dynamic changes in EEG signals across various time scales through multi-scale grouped convolution and backbone temporal convolution skip connections; the latter improves local spatial feature extraction and calibrates feature mapping through the introduction of cross-domain spatial filtering layers. Furthermore, by specifically optimizing the loss function, MFRC-Net effectively reduces sensitivity to outliers. Experiment results on the BCI Competition IV 2a dataset and the SHU dataset demonstrate that, with a parameter size of only 13 K, MFRC-Net achieves accuracy of 85.1% and 69.3%, respectively, surpassing current state-of-the-art models. The integration of temporal multi-scale residual convolution blocks and cross-domain dual-stream spatial convolution blocks in lightweight models significantly boosts performance, as evidenced by ablation studies and visualizations.

Active suspended backpacks represent a promising solution to mitigate the impact of inertial forces on individuals engaged in load carriage. However, identifying effective control objectives aimed at enhancing human carrying capacity remains a significant challenge. In this study, we introduce a novel approach by integrating a limb-like structure-type (LLS) bioinspired vibration isolator, modeled using Lagrangian mechanics, into an active load-transfer suspended backpack to primarily alleviate human shoulder pressure, thereby constructing a humanrobot interaction control framework for the system. Drawing from a double-mass coupled oscillator model, this approach formulates a vertical dynamics model for the human-backpack system, systematically exploring the principles of both static load transfer and dynamic load reduction on the human shoulder. Subsequently, a series elastic actuators-based controller with prescribed performance is proposed to simultaneously achieve trajectory tracking and ensure load motion within the limited range. Theoretically, we validate the input-output stability of the LLS model and guarantee the ultimate uniform boundedness of the closed-loop system. Simulation and experimental trials conducted across different terrain scenarios validate the effectiveness of the proposed method, highlighting reductions of 18.68% in metabolic rate during level ground walking, 9.58% in a staircase scenario and 12.35% in a complex terrain, involving uphill, downstairs, and flat ground walking.

Supernumerary robotic limbs (SRLs) offer substantial potential in both the rehabilitation of hemiplegic patients and the enhancement of functional capabilities for healthy individuals. Designing a general-purpose SRL device is inherently challenging, particularly when developing a unified theoretical framework that meets the diverse functional requirements of both upper and lower limbs. In this paper, we propose a multi-objective optimization (MOO) design theory that integrates grasping workspace similarity, walking workspace similarity, braced force for sit-to-stand (STS) movements, and overall mass and inertia. A geometric vector quantification method is developed using an ellipsoid to represent the workspace, aiming to reduce computational complexity and address quantification challenges. The ellipsoid envelope transforms workspace points into ellipsoid attributes, providing a parametric description of the workspace. Furthermore, the STS static braced force assesses the effectiveness of force transmission. The overall mass and inertia restricts excessive link length. To facilitate rapid and stable convergence of the model to high-dimensional irregular Pareto fronts, we introduce a multi-subpopulation correction firefly algorithm. This algorithm incorporates a strategy involving attractive and repulsive domains to effectively handle the MOO task. The optimized solution is utilized to redesign the prototype for experimentation to meet specified requirements. Six healthy participants and two hemiplegia patients participated in real experiments. Compared to the pre-optimization results, the average grasp success rate improved by 7.2%, while the muscle activity during walking and STS tasks decreased by an average of 12.7% and 25.1%, respectively. The proposed design theory offers an efficient option for the design of multi-functional SRL mechanisms.

Lower-limb motion intent recognition is a crucial aspect of wearable robot control and human-machine collaboration. Among the various sensors used for this purpose, the electromyogram (EMG) sensor remains one of the most widely employed. However, EMG signals are highly susceptible to electrical noise, motion artefacts, and perspiration, which can compromise their quality. To address these challenges, we designed an air-pressure mechanomyography (PMMG) sensor and developed a wearable multi-modal sensor system that incorporates PMMG thigh-ring, inertial measurement unit (IMU), and force-sensitive resistor (FSR). To enhance gait phase and locomotion mode recognition performance, we proposed a gate multi-task TSK fuzzy inference system (GMT-TSK-FIS) algorithm that enables simultaneous handling of multiple recognition tasks. This approach enabled the development of a lower-limb motion intent recognition system that can simultaneously recognize gait phase and locomotion mode based on GMT-TSK-FIS. The experimental results showed that the accuracy of gait phase and locomotion mode recognition was 98.28% and 99.96%, respectively. Furthermore, the study demonstrated that multi-modal sensor fusion outperformed single-modal sensor fusion, while multi-task recognition exhibited better performance than single-task recognition.

This paper investigates the stability problem of Takagi-Sugeno (T-S) fuzzy systems with time-varying delay. Different from some existing methods, a new delay-productdependent Lyapunov-Krasovkii functional (LKF) is proposed, whose derivative is estimated by using the proposed cubic function negative-determination lemma (NDL). Moreover, a parameter-dependent reciprocally convex inequality (PDRCI) is proposed to improve the estimation accuracy of reciprocally convex terms. Besides, an improved cubic function NDL is derived to solve the negative-definiteness determination of cubic functions. Based on the proposed delay-product-dependent LKF, the developed PDRCI and the improved cubic function NDL, a less conservative stability criterion is obtained. Three examples are presented to demonstrate the merits of the proposed approaches by comparing with several existing results.

In this article, considering three different working modes (including the human-supervised, human-aided, and human-manned modes), we present a novel mode-switched control architecture for solving the human-in-the-loop (HIL) teleoperation problem of multislave robots with local slave-to-slave (S2S) communication, long-distance master-to-slave (M2S) communication and transmission delays. Throughout the control process, the S2S communication is updated following the event-triggered mechanism; meanwhile, the data transmission is fully distributed, namely, no global information can be transmitted. Besides, we also deal with the concerns of enhancing “telepresence”, namely, reconstructing the interaction force between a user-determined slave robot and its task environment at the human side, and then allow the human operator to control the multislave robots in a “virtual reality” way. To this end, by making full use of the historical and current data, the data-training-based (DTB) observer is designed to obtain the interaction force at the side of the slave robot and then assist the human operator to choose a proper control mode. The presented architecture is hierarchically designed for data collection, data processing and physical regulation, involving the DTB observer and the fully distributed event-based (FDEB) estimator in a unified framework. Finally, numerical examples are conducted to demonstrate the effectiveness of the architecture.

Bootstrap aggregating (Bagging) and boosting are two popular ensemble learning approaches, which combine multiple base learners to generate a composite model for more accurate and more reliable performance. They have been widely used in biology, engineering, healthcare, etc. This article proposes BoostForest, which is an ensemble learning approach using BoostTree as base learners and can be used for both classification and regression. BoostTree constructs a tree model by gradient boosting. It increases the randomness (diversity) by drawing the cut-points randomly at node splitting. BoostForest further increases the randomness by bootstrapping the training data in constructing different BoostTrees. BoostForest generally outperformed four classical ensemble learning approaches (Random Forest, Extra-Trees, XGBoost and LightGBM) on 35 classification and regression datasets. Remarkably, BoostForest tunes its parameters by simply sampling them randomly from a parameter pool, which can be easily specified, and its ensemble learning framework can also be used to combine many other base learners.

Objective:While neuroscience research has established a link between vision and intention, studies on gaze data features for intention recognition are absent. The majority of existing gaze-based intention recognition approaches are based on deliberate long-term fixation and suffer from insufficient accuracy. In order to address the lack of features and insufficient accuracy in previous studies, the primary objective of this study is to suppress noise from human gaze data and extract useful features for recognizing grasp intention.Methods:We conduct gaze movement evaluation experiments to investigate the characteristics of gaze motion. The target-attracted gaze movement model (TAGMM) is proposed as a quantitative description of gaze movement based on the findings. A Kalman filter (KF) is used to reduce the noise in the gaze data based on TAGMM. We conduct gaze-based natural grasp intention recognition evaluation experiments to collect the subject's gaze data. Four types of features describing gaze point dispersion ($f_{var}$), gaze point movement ($f_{gm}$), head movement ($f_{hm}$), and distance from the gaze points to objects ($f_{d_{j}}$) are then proposed to recognize the subject's grasp intentions. With the proposed features, we perform intention recognition experiments, employing various classifiers, and the results are compared with different methods.Results:The statistical analysis reveals that the proposed features differ significantly across intentions, offering the possibility of employing these features to recognize grasp intentions. We demonstrated the intention recognition performance utilizing the TAGMM and the proposed features in within-subject and cross-subject experiments. The results indicate that the proposed method can recognize the intention with accuracy improvements of 44.26% (within-subject) and 30.67% (cross-subject) over the fixation-based method. The proposed method also consumes less time (34.87 ms) to recognize the intention than the fixation-based method (about 1 s).Conclusion:This work introduces a novel TAGMM for modeling gaze movement and a variety of practical features for recognizing grasp intentions. Experiments confirm the effectiveness of our approach.Significance:The proposed TAGMM is capable of modeling gaze movements and can be utilized to process gaze data, and the proposed features can reveal the user's intentions. These results contribute to the development of gaze-based human-robot interaction.

This article proposes a distributed cooperative manipulation control scheme for multirobot systems to track reference trajectories with unknown payload dynamics, grasp positions, and external disturbances. An online learning module is established to estimate the payload dynamics. Then a wrench-synthetic trajectory tracking control protocol is thereby developed to manipulate an object under unknown external disturbances no matter where the grasping points are. Moreover, sufficient conditions are derived to guarantee the uniform boundedness of the tracking errors of the closed-loop cooperative manipulation system. Finally, numerical simulations are conducted to substantiate the effectiveness of the proposed cooperative manipulation control scheme.

The dynamic load attached to the load gravity imposes an excessive burden to human shoulders during load carriage, resulting in possible muscle injuries and additional physical exertion. This paper proposes an active suspension backpack, capable of transferring partial load from human shoulders to pelvis and alleviating the dynamic load through separated panels and motor actuation, to reduce pressure on human shoulders and improve walking metabolic efficiency. Based on the human body motion in the vertical direction, the dynamical model of the human-backpack system with shoulder interaction force measured by a soft ballonet with an embedded air pressure sensor is introduced, and an impedance controller has been implemented to maintain a relatively small and constant pressure on the shoulder. In an experimental case study, we presents preliminary results of three healthy subjects performing a treadmill walking with a 20kg load in ACTIVE configuration where the shoulder pressure shows a decrease by 30% along with a reduction of the metabolic energy consumption by 16.4%, compared with the load LOCKED case.

Although the grasping state analysis is vital in the study of manipulators, the grasping state analysis of soft manipulators as an independent research topic is not much so far. This paper proposes a novel pneumatic soft manipulator with a flexible tactile sensor array (SM-FTSA). The flexible tactile sensor array comprises piezoresistive materials with a porous structure. An equal potential approach is adopted to realize the collection of tactile signals of the SM-FTSA. Inspired by the grasping analysis of rigid manipulators, we propose 4 grasping states for the SM-FTSA, including inflating, shaking, stable, and slipping. Based on the experimental data, we conduct grasping experiments on 12 objects with SM-FTSA, and we propose 10 features that reflect the grasping state. Several machine learning methods are utilized to classify the grasping state. Among them, the Random Forest method presents the best performance, and the average classification accuracy reaches 99%.

Acquisition of labeled training samples for affective computing is usually costly and time-consuming, as affects are intrinsically subjective, subtle and uncertain, and hence multiple human assessors are needed to evaluate each affective sample. Particularly, for affect estimation in the 3D space of valence, arousal and dominance, each assessor has to perform the evaluations in three dimensions, which makes the labeling problem even more challenging. Many sophisticated machine learning approaches have been proposed to reduce the data labeling requirement in various other domains, but so far few have considered affective computing. This paper proposes two multi-task active learning for regression approaches, which select the most beneficial samples to label, by considering the three affect primitives simultaneously. Experimental results on the VAM corpus demonstrated that our optimal sample selection approaches can result in better estimation performance than random selection and several traditional single-task active learning approaches. Thus, they can help alleviate the data labeling problem in affective computing, i.e., better estimation performance can be obtained from fewer labeling queries.

The papers in this special section focus on cyborg intelligence. Well-known scientists and experts have expressed concern that robots may take over the world. More generally, there is a concern that robots could take over human jobs and leave billions of people suffering long-term unemployment. Yet, such concerns ignored the potential of intelligence techniques to enhance the natural capabilities of human beings with in-the-body technologies and so become cyborgs with superior capabilities to robots. Cyborg intelligence is dedicated to improving the natural capabilities of human beings by integrating artificial intelligence (AI) with biological intelligence and in-the-body technologies through tight integrations of machines and biological beings.

As an important movement of the daily living activities, sit-to-stand (STS) movement is usually a difficult task facing elderly and dependent people. In this article, a novel impedance modulation strategy of a lower-limb exoskeleton is proposed to provide appropriate power and balance assistance during STS movements while preserving the wearer’s control priority. The impedance modulation control strategy ensures adaptation of the mechanical impedance of the human–exoskeleton system toward a desired one requiring less wearer’s effect while reinforcing the wearer’s balance control ability during STS movements. A human joint torque observer is designed to estimate the joint torques developed by the wearer using joint position kinematics instead of electromyography or force sensors; a time-varying desired impedance model is proposed according to the wearer’s lower-limb motion ability. A virtual environmental force is designed for balance reinforcement control. Stability and robustness of the proposed method are theoretically analyzed. Simulations are implemented to illustrate the characteristics and performance of the proposed approach. Experiments with four healthy subjects are carried out to evaluate the effectiveness of the proposed method and show satisfactory results in terms of appropriate power assist and balance reinforcement.

Dimensionality reduction is an important operation in information visualization, feature extraction, clustering, regression, and classification, especially for processing noisy high dimensional data. However, most existing approaches preserve either the global or the local structure of the data, but not both. Approaches that preserve only the global data structure, such as principal component analysis (PCA), are usually sensitive to outliers. Approaches that preserve only the local data structure, such as locality preserving projections, are usually unsupervised (and hence cannot use label information) and uses a fixed similarity graph. We propose a novel linear dimensionality reduction approach, supervised discriminative sparse PCA with adaptive neighbors (SDSPCAAN), to integrate neighborhood-free supervised discriminative sparse PCA and projected clustering with adaptive neighbors. As a result, both global and local data structures, as well as the label information, are used for better dimensionality reduction. Classification experiments on nine high-dimensional datasets validated the effectiveness and robustness of our proposed SDSPCAAN.

Takagi-Sugeno-Kang (TSK) fuzzy systems are flexible and interpretable machine learning models; however, they may not be easily optimized when the data size is large, and/or the data dimensionality is high. This article proposes a minibatch gradient descent (MBGD) based algorithm to efficiently and effectively train TSK fuzzy classifiers. It integrates two novel techniques: First, uniform regularization (UR), which forces the rules to have similar average contributions to the output, and hence to increase the generalization performance of the TSK classifier; and, second, batch normalization (BN), which extends BN from deep neural networks to TSK fuzzy classifiers to expedite the convergence and improve the generalization performance. Experiments on 12 UCI datasets from various application domains, with varying size and dimensionality, demonstrated that UR and BN are effective individually, and integrating them can further improve the classification performance.

Fuzzy systems have achieved great success in numerous applications. However, there are still many challenges in designing an optimal fuzzy system, e.g., how to efficiently optimize its parameters, how to balance the trade-off between cooperations and competitions among the rules, how to overcome the curse of dimensionality, how to increase its generalization ability, etc. Literature has shown that by making appropriate connections between fuzzy systems and other machine learning approaches, good practices from other domains may be used to improve the fuzzy systems, and vice versa. This article gives an overview on the functional equivalence between Takagi-Sugeno-Kang fuzzy systems and four classic machine learning approaches-neural networks, mixture of experts, classification and regression trees, and stacking ensemble regression-for regression problems. We also point out some promising new research directions, inspired by the functional equivalence, that could lead to solutions to the aforementioned problems. To our knowledge, this is so far the most comprehensive overview on the connections between fuzzy systems and other popular machine learning approaches, and hopefully will stimulate more hybridization between different machine learning algorithms.

Takagi–Sugeno–Kang (TSK) fuzzy systems are very useful machine learning models for regression problems. However, to our knowledge, there has not existed an efficient and effective training algorithm that ensures their generalization performance and also enables them to deal with big data. Inspired by the connections between TSK fuzzy systems and neural networks, we extend three powerful neural network optimization techniques, i.e., minibatch gradient descent (MBGD), regularization, and AdaBound, to TSK fuzzy systems, and also propose three novel techniques (DropRule, DropMF, and DropMembership) specifically for training TSK fuzzy systems. Our final algorithm, MBGD with regularization, DropRule, and AdaBound, can achieve fast convergence in training TSK fuzzy systems, and also superior generalization performance in testing. It can be used for training TSK fuzzy systems on datasets of any size; however, it is particularly useful for big datasets, on which currently no other efficient training algorithms exist.

Pneumatic muscle actuators (PMAs), a kind of soft/compliant actuators, have been attracted a great deal of attention in the studies of rehabilitation robots. However, the nonlinearities, uncertainties, hysteresis, and time-varying features of PMAs bring a lot of difficulties in their high-precision trajectory tracking tasks. In this paper, an echo state Gaussian process-based nonlinear model predictive control (ESGP-NMPC) is designed for the PMAs. The proposed strategy is comprised of an ESGP, which is suitable for modeling unknown nonlinear systems as well as measuring their uncertainties, and a gradient descent optimization algorithm for calculating the control signal sequences. Based on the Lyapunov theorem, characteristics of the closed-loop system are analyzed to guarantee the asymptotical stability. Both simulations and physical experiments are carried out to illustrate the validity of the proposed control strategy. Compared with other conventional methods, the ESGP-NMPC can achieve a better model fitting for the PMA and control performance for the high-precision tracking tasks. Note to Practitioners-High-precision control of pneumatic muscle actuators (PMAs) is a vital problem when PMAs are utilized as actuators of rehabilitation robots since the patient's safety and the performance of rehabilitation tasks are largely dependent on the accuracy of the actuators. Conventional model-based control approaches usually require relatively accurate identification of system parameters, which is difficult for the PMA, owing to its strong nonlinear and time-varying characteristics. This paper proposes a new model predictive control method based on an echo state Gaussian process that can describe the unknown dynamics of a PMA due to its universal approximation property. Through the optimization method, the controller can be efficiently realized and presents better performances than some comparatives. By applying this approach, it is possible to achieve not only high-precision control of PMAs but also a certain degree of robustness to the load.

The representation theorem for interval type-2 fuzzy sets (IT2 FSs), proposed by Mendel and John, states that an IT2 FS is a combination of all its embedded type-1 (T1) FSs, which can be nonconvex and/or subnormal. These nonconvex and/or subnormal embedded T1 FSs are included in developing many theoretical results for IT2 FSs, including uncertainty measures, the linguistic weighted averages (LWAs), the ordered LWAs (OLWAs), the linguistic weighted power means (LWPMs), etc. However, convex and normal T1 FSs are used in most fuzzy logic applications, particularly computing with words. In this paper, we propose a constrained representation theorem (CRT) for well-shaped IT2 FSs using only its convex and normal embedded T1 FSs, and show that IT2 FSs generated from three word encoding approaches and four computing with words engines (LWAs, OLWAs, LWPMs, and perceptual reasoning) are all well-shaped IT2 FSs. We also compute five constrained uncertainty measures (centroid, cardinality, fuzziness, variance, and skewness) for well-shaped IT2 FSs using the CRT. The CRT and the associated constrained uncertainty measures can be useful in computing with words, IT2 fuzzy logic system design using the principles of uncertainty, and measuring the similarity between two well-shaped IT2 FSs.

This paper addresses the fault detection and isolation (FDI) problem for robotic assembly of electrical connectors in the framework of set-membership. Both the fault-free and faulty cases of assembly are modeled by different switched linear models with known switching sequences, bounded parameters, and external disturbances. The locations of switching points of each model are assumed to be inside some areas but the accurate positions are not clear. Given current input/output data, the feasible parameter set of fault-free switched linear model is obtained by sequentially calculating an optimal ellipsoid. If the pair of data is not consistent with any possible submodel, a fault is then detected. The isolation of fault is realized by checking the consistency between the data sequence and each possible fault model one by one. The robustness of the proposed FDI algorithms is proved. The effectiveness of these algorithms is verified by the robotic assembly experiments of mating electrical connectors.

This paper presents an integrated interval type-2 fuzzy logic approach that simultaneously models and controls an underactuated mobile two-wheeled inverted pendulum (MTWIP), which suffers from modeling uncertainties and external disturbances. The control objective is to attain the desired position and direction while keeping the MTWIP balanced. It is achieved by integrating four interval type-2 fuzzy logic systems (IT2 FLSs): the first IT2 FLS describes the dynamics of the MTWIP using a Takagi-Sugeno model, the second IT2 FLS controls the balance of the MTWIP using also a Takagi-Sugeno model, and the third and fourth IT2 FLSs control its position and direction, respectively, using a Mamdani model. A linear matrix inequality based design approach is also proposed to guarantee the stability of the balance controller. The proposed approach is compared with a type-1 FLS in real-world experiments. All results demonstrate that the IT2 FLS outperforms the type-1 FLS, especially under modeling uncertainties and external disturbances.

A wearable, portable, low-cost, and easy-to-use upper extremity exoskeleton robot, RUPERT, is presented here for clinical and in-home therapies of patients who have survived a stroke. The robot system has five degrees-of-freedom and is driven by compliant and safe pneumatic muscles. Its primary function is to assist the movement of an affected arm in 3-D space and perform a daily training program in a virtual environment. Subjects are recruited for experiments of both clinical and in-home therapies. While using RUPERT, most subjects exhibited significant improvements in various functional measures. During their in-home therapy trials, subjects expressed enthusiasm regarding the modules and the use of our robotic system. The experimental results look promising and the proposed robotic system exhibits a good prospect for a future commercialized in-home therapy product.

Fall prevention is one of the most important functions of walking assistance devices for user's safety. It is preferable that these devices obviates factors which induce falling over rather than helping them recovering from falling motion. Tandem stance, where both legs form a line along walking direction, is a factor of falling. It is often observed in turning motion because a swing leg moves in lateral direction as well as forward. Generally, upper body's turning of a person proceeds before lower limb's turning during walking, and the behavior of the lower limb is constrained by upper body turning. It is thus possible to control behavior of the lower limb for tandem stance prevention by constraining behavior of the upper body. This paper therefore introduces a tandem stance prevention method for safer walk of the elderly or physically challenged person. The method adjusts admittance of cane robot's turning motion according to the positions of a support leg and a swing leg. The cane robot measures user's leg position and estimates a walk phase of the user. By adjusting an admittance model of the cane robot based on the walk phase, the cane robot resists to turn while support leg is on the same side of the turning direction. As a result, the tandem stance is avoided. Through experiments, we confirmed that the cane robot successfully controls user's upper body's behavior to prevent the user from being in tandem stance.

A cane-type robot called intelligent cane has been developed to support the elderly during walking. By supporting a part of a user's body weight, the cane robot aims to reduce a load applied to a user's affected leg. Therefore, while the user's affected leg is a support leg, it is preferable that the cane robot stops to sufficiently support the user. In our previous work, the cane robot is controlled based on horizontal component of force applied to the cane robot and moment around a vertical axis. In this paper, virtual friction force, which is proportional to vertical component of force, is proposed to improve a walking assistance capability of the cane robot. In addition, virtual frictional coefficients are arranged based on the user's state inferred by a laser range finder. By employing the proposed method, the cane robot moves easily in the both legs support phase, stops in the healthy leg support phase, and supports the user reliably in the affected leg support phase.

Recognition of the wearer's motion intention plays an important role in the study of power-assist robots. In this paper, an intention-guided control strategy is proposed and applied to an upper-limb power-assist exoskeleton. Meanwhile, a human-robot interface comprised of force-sensing resistors (FSRs) is designed to estimate the motion intention of the wearer's upper limb in real time. Moreover, a new concept called the “intentional reaching direction (IRD)” is proposed to quantitatively describe this intention. Both the state model and the observation model of IRD are obtained by studying the upper limb behavior modes and analyzing the relationship between the measured force signals and the motion intention. Based on these two models, the IRD can be inferred online using an adapted filtering technique. Guided by the inferred IRD, an admittance control strategy is deployed to control the motions of three DC motors placed at the corresponding joints of the robotic arm. The effectiveness of the proposed approaches is finally confirmed by experiments on a 3 degree-of-freedom (DOF) upper-limb robotic exoskeleton.

Fault detection functions with learning method of a robotic manipulator are very useful for factory automation. All production has the possibility to fail due to unexpected accidents. To reduce the fatigue of human workers, small errors automatically should be corrected by a robot system. Also a learning method is important for fault detection, because labor of system integrator should be reduced. In this paper, an external thread fastening task by a robotic manipulator is investigated. To discriminate the four states of a task, linear support vector machine methods with two feature parameters are introduced. The effectiveness of the proposed algorithm is confirmed through an experiment and recognition examination. Finally, the ability of linear SVM is compared with artificial neural network method.

Intelligent cane is a robot developed for assisting the elderly or handicapped people in walking. As a nursing-care device, the cane robot is designed to assist the elderly or handicapped people not only in indoor environments but also in outdoor environments. Therefore the cane robot is required to be smaller and lighter. In addition, it is preferred that the cane robot is movable in omni-directions so that it can be used in various situations. A concept called “intentional direction (ITD)” was proposed to estimate the user's walking intention by analyzing signals from a 6-axis force/torque sensor fixed to the handle of the aluminum stick. Admittance control method was applied to the motion control of the cane robot. In this paper, a new algorithm based on the usage and purpose of an ordinary cane is proposed. In the proposed algorithm, the cane robot is appropriately stopped to support the elderly more effectively. The effectiveness of the proposed method is verified through the experiments.

In electronic manufacturing system, the design of the robotic gripper is important for the successful accomplishment of the assembly task. Due to the restriction of the architecture of traditional robotic hands, the status of assembly parts during the assembly process cannot be effectively detected. In this research, an intelligent robotic gripper - i-Hand equipped with multiple small sensors is designed and built for this purpose, getting the essential parameters for some specific mathematical model. Mating connectors by robot, as an experimental case in this paper, is studied to evaluate the performance of i-Hand. A simple new model is proposed to describe the process of mating connectors, within which the distance between the connector and deformable Printed Circuit Board (PCB) is detected by i-Hand. An online Fault Detection and Diagnosis (FDD) algorithm is proposed. Various possible situations during assembly are considered and handled according to an event driven work flow. The effectiveness of proposed model and algorithm is proved by the experiments.

The traffic congestion is growing in urban area these days due to the rapidly increasing number of automobiles. One of the traffic problem solutions is to increase the flow rate of a particular traffic artery. To make the vehicle smaller and narrower might be a good idea. We developed a new conceptual Narrow Tilting Vehicle (NTV) that has one front wheel and two rear wheels. The proposed NTV is very narrow and all three wheels can tilt to improve the stability during turning. We also derive a new switching dynamical model of the NTV. This model considers several states including normal running, temporary running with one rear wheel not on the ground and totally falling down. Based on this model, a simulation platform is established, which is useful to test different control methods and to investigate dangerous driving situations in case hurt drivers in real experiments. The effectiveness of proposed dynamical model is verified through comparison study of simulations and corresponding experiments.

An intelligent cane robot (iCane) was designed for aiding the elderly who have muscle weakness on lower limbs. A commercial omni-directional wheels robot was used as an omni-directional mobile base, and an aluminum stick was installed on the base of cane robot. A Concept called “intentional direction (ITD)” was proposed for estimating the user's walking intention by analyzing the signal of a 6-axis force/torque sensor which is fixed to the handle of stick. A universal joint driven by two DC motors was designed to control the posture of the stick. As a care-nursing device, the cane robot was designed to assist the elderly in both indoor and outdoor environments. Therefore the size and weight of cane robot should be minimized. But in that case, there is high risk that the cane robot would be pushed over by the user. In this paper a constrained nonlinear multivariable algorithm was designed to optimize the stable posture of cane robot. By controlling the posture of stick, the maximums sufferable torque moment which lead to cane robot falling over can be increased. The experimental results show that the stability of cane robot can be enhanced effectively.

Recognizing the user motion intention plays an important role in the study of power-assist robots. An intention-guided control strategy is proposed for the upper-limb power-assist exoskeleton. A force sensor system comprised of force sensing resistors (FSRs) is designed to online estimate the motion intention of user upper limb. A new concept called "intentional reaching direction (IRD)" is proposed to quantitatively describe this intention. Both the state model and the observation model of IRD are obtained by enumerating the upper limb behavior modes and analyzing the relationship between the measured force signals and the motion intention. Based on these two models, the IRD can be online inferred by applying filtering technology. Guided by the estimated IRD, an admittance control strategy is assumed to control the motions of three DC motors in the joints of the robotic arm. The effectiveness of the proposed approaches is finally confirmed by the experiments on a 3-DOF robotic exoskeleton.

Manufacturing industry tends to employ more flexible assembly cells for High-Mix, Low-Volume production. We has proposed an innovative human and robot hybrid assembly cell within this purpose to solve the problem of persistent growing cost for human resources and now and again changes in programs and configurations for robots, and to achieve a high manufacturing efficiency. One of the key issues is to find out the optimal way of allocating the assembly subtasks to both human and robot. In this paper, a model for assembly strategy generation and selection for human and robot coordinated (HRC) cell assembly is proposed. A Dual Generalized Stochastic Petri Net (GSPN) model is theoretically researched and then built based on a practical assembly task for human and robot coordination. Based on GSPN, Monte Carlo method is carried out to study the time cost and payment cost for possible strategies, and Multiple-Objective Optimization (MOOP) method related Cost-effectiveness analysis is adopted to select the optimal ones. We demonstrate the effectiveness of this approach by comparing the simulation and experimental results.

In this study, we present an omni-directional cane robot for aiding the elderly and handicapped people walking. The motion control problem is investigated for both normal and abnormal walking conditions. In the case of the user's normal walking aided by the cane robot, a concept called “Intentional Direction (ITD)” is proposed. Guided by the online estimated ITD, we apply the admittance control method in the motion control of cane robot. The gravity compensation is also considered because the user is possible to walk on a slope. For the abnormal walking, we mainly studied the case of user's falling down. The Center of Gravity (COG) of user can be estimated from the angle of an inverted pendulum which represents human model. Fall prevention algorithm based on the relationship between user's COG and the cane is proposed. This algorithm is also applied in the motion control of cane robot. The proposed method are verified through experiments.

There has been increasing interest in a type of underactuated mechanical systems, mobile-wheeled inverted-pendulum (MWIP) models, which are widely used in the field of autonomous robotics and intelligent vehicles. Robust-velocity-tracking problem of the MWIP systems is investigated in this study. In the velocity-control problem, model uncertainties accompany uncertain equilibriums, which make the controller design become more difficult. Two sliding-mode-control (SMC) methods are proposed for the systems, both of which are capable of handling both parameter uncertainties and external disturbances. The asymptotical stabilities of the corresponding closed-loop systems are achieved through the selection of sliding-surface parameters, which are based on some rules. There is still a steady tracking error when the first SMC controller is used. By assuming a novel sliding surface, the second SMC controller is designed to solve this problem. The effectiveness of the proposed methods is finally confirmed by the numerical simulations.

There has been an increasing interest in a kind of underactuated mechanical systems, mobile wheeled inverted pendulum (MWIP) models, which are widely used in the field of autonomous robotics and intelligent vehicles. Robust velocity tracking problem of MWIP systems is investigated in this study. In the velocity control problem, model uncertainties accompany uncertain equilibriums, which make the controller design become more difficult. A sliding mode control (SMC) method based on a novel sliding surface is proposed for the systems, which are capable of handling both parameter uncertainties and external disturbances. By assuming the specially designed sliding surface, the proposed SMC controller is capable of eliminating the steady velocity tracking error. The asymptotical stability of the closed-loop system is achieved through selecting sliding surface parameters in terms of some rules. The effectiveness of the proposed methods is finally confirmed by numerical simulations.

A three-wheeled omni-directional cane robot is designed for aiding the elderly walking. Possible move modes are analyzed and a corresponding hybrid model is constructed to describe walking behavior. A concept called intentional direction is presented to denote the moving intention of a human. Based on experiments and some assumptions, dynamic model of intentional direction is obtained as well as its online estimation method. A new admittance control scheme is presented based on intentional direction, which provides natural and intuitive human machine interface. Experiment results show the effectiveness of the design.