Glove-Net: Enhancing Grasp Classification with Multisensory Data and Deep Learning Approach.

Grasp classification is pivotal for understanding human interactions with objects, with wide-ranging applications in robotics, prosthetics, and rehabilitation. This study introduces a novel methodology utilizing a multisensory data glove to capture intricate grasp dynamics, including finger posture bending angles and fingertip forces. Our dataset comprises data collected from 10 participants engaging in grasp trials with 24 objects using the YCB object set. We evaluate classification performance under three scenarios: utilizing grasp posture alone, utilizing grasp force alone, and combining both modalities. We propose Glove-Net, a hybrid CNN-BiLSTM architecture for classifying grasp patterns within our dataset, aiming to harness the unique advantages offered by both CNNs and BiLSTM networks. This model seamlessly integrates CNNs' spatial feature extraction capabilities with the temporal sequence learning strengths inherent in BiLSTM networks, effectively addressing the intricate dependencies present within our grasping data. Our study includes findings from an extensive ablation study aimed at optimizing model configurations and hyperparameters. We quantify and compare the classification accuracy across these scenarios: CNN achieved 88.09%, 69.38%, and 93.51% testing accuracies for posture-only, force-only, and combined data, respectively. LSTM exhibited accuracies of 86.02%, 70.52%, and 92.19% for the same scenarios. Notably, the hybrid CNN-BiLSTM proposed model demonstrated superior performance with accuracies of 90.83%, 73.12%, and 98.75% across the respective scenarios. Through rigorous numerical experimentation, our results underscore the significance of multimodal grasp classification and highlight the efficacy of the proposed hybrid Glove-Net architectures in leveraging multisensory data for precise grasp recognition. These insights advance understanding of human-machine interaction and hold promise for diverse real-world applications.

Design and validation of a pediatric gait assistance exoskeleton system with fast non-singular terminal sliding mode controller.

Existing exoskeletons for pediatric gait assistance have limitations in anthropometric design, structure weight, cost, user safety features, and adaptability to diverse users. Additionally, creating precise models for pediatric rehabilitation is difficult because the rapid anthropometric changes in children result in unknown model parameters. Furthermore, external disruptions, like unpredictable movements and involuntary muscle contractions, add complexity to the control schemes that need to be managed. To overcome these limitations, this study aims to develop an affordable stand-aided lower-limb exoskeleton specifically for pediatric subjects (8-12 years, 25-40 kg, 128-132 cm) in passive-assist mode. The authors modified a previously developed model (LLESv1) for improved rigidity, reduced mass, simplified motor arrangement, variable waist size, and enhanced mobility. A computer-aided design of the new exoskeleton system (LLESv2) is presented. The developed prototype of the exoskeleton appended with a pediatric subject (age: 12 years old, body mass: 40 kg, body height: 132 cm) is presented with real-time hardware architecture. Thereafter, an improved fast non-singular terminal sliding mode (IFNSTSM) control scheme is proposed, incorporating a double exponential reaching law for expedited error convergence and enhanced stability. The Lyapunov stability warrants the control system's performance despite uncertainties and disturbances. In contrast to fast non-singular terminal sliding mode (FNSTSM) control and time-scaling sliding mode (TSSM) control, experimental validation demonstrates the effectiveness of IFNSTSM control by a respective average of 5.39% and 42.1% in tracking desired joint trajectories with minimal and rapid finite time converging errors. Moreover, the exoskeleton with the proposed IFNSTSM control requires significantly lesser control efforts than the exoskeleton using contrast FNSTSM control. The Bland-Altman analysis indicates that although there is a minimal mean difference in variables when employing FNSTSM and IFNSTSM controllers, the latter exhibits significant performance variations as the mean of variables changes. This research contributes to affordable and effective pediatric gait assistance, improving rehabilitation outcomes and enhancing mobility support.

Robust LQR-Based Neural-Fuzzy Tracking Control for a Lower Limb Exoskeleton System with Parametric Uncertainties and External Disturbances.

The design of an accurate control scheme for a lower limb exoskeleton system has few challenges due to the uncertain dynamics and the unintended subject's reflexes during gait rehabilitation. In this work, a robust linear quadratic regulator- (LQR-) based neural-fuzzy (NF) control scheme is proposed to address the effect of payload uncertainties and external disturbances during passive-assist gait training. Initially, the Euler-Lagrange principle-based nonlinear dynamic relations are established for the coupled system. The input-output feedback linearization approach is used to transform the nonlinear relations into a linearized state-space form. The architecture of the adaptive neuro-fuzzy inference system (ANFIS) and used membership function are briefly explained. While varying mass parameters up to 20%, three robust neural-fuzzy datasets are formulated offline with the joint error vector and LQR control input. Thereafter, to deal with external interferences, an error dynamics with a disturbance estimator is presented using an online adaptation of the firing strength matrix. The Lyapunov theory is carried out to ensure the asymptotic stability of the coupled human-exoskeleton system in view of the proposed controller. The gait tracking results for the proposed control scheme (RLQR-NF) are presented and compared with the exponential reaching law-based sliding mode (ERL-SM) controller. Furthermore, to investigate the robustness of the proposed control over LQR control, a comparative performance analysis is presented for two cases of parametric uncertainties and external disturbances. The first case considers the 20% raise in mass values with a trigonometric form of disturbances, and the second case includes the effect of the 30% increment in mass values with a random form of disturbances. The simulation runs have shown the promising gait tracking aspects of the designed controller for passive-assist gait training.

Preliminary design and development of a low-cost lower-limb exoskeleton system for paediatric rehabilitation.

In this work, the design, modeling, and development of a low-cost lower limb exoskeleton (LLES) system are presented for paediatric rehabilitation (age: 8-12 years, mass: 25-40 kg, height: 115-125 cm). The exoskeleton system, having three degrees-of-freedom (DOFs) for each limb, is designed in the SolidWorks software. A wheel support module is introduced in the design to ensure the user's stability and safety. The finite element analysis of the hip joint connector along with the wheel support module is realized for maximum loading conditions. The holding torque capacity of exoskeleton joints is estimated using an affordable spring-based experimental setup. A working prototype of the LLES is developed with holding torque rated actuators. Thereafter, the dynamic analysis for the human-exoskeleton coupled system is carried out using the Euler-Lagrange principle and SimMechanics model. The simulation results of estimating joint actuator torques are obtained for two paraplegic subjects (Case I: 10 years age, 30 kg mass, 120 cm height and Case II: 12 years age, 40 kg mass, 125 cm height). The details of input parameters such as body mass, link lengths, joint angles, and contact forces are discussed. The simulation results of dynamic analysis have shown the potential of estimating the torques of joint actuators for the developed prototype during motion assistance and gait rehabilitation.

Adaptive neuro-fuzzy inference system-based path planning of 5-degrees-of-freedom spatial manipulator for medical applications.

Over the last few decades, medical-assisted robots have been considered by many researchers, within the research domain of robotics. In this article, a 5-degrees-of-freedom spatial medical manipulator is analyzed for path planning, based on inverse kinematic solutions. Analytical methods have generally employed for finding the inverse kinematic solutions in earlier studies. However, this method is only appreciable in case of closed-form solutions. The unusual joint configurations of considered manipulator result in more complexity to attain the closed-form solutions, analytically. To overcome with shortcomings of analytical method, a non-traditional approach named adaptive neuro-fuzzy inference system is proposed under the class of artificial intelligent techniques. This article presents this neuro-fuzzy approach for desired path generation by 5-degrees-of-freedom manipulator. The estimation of percentage error between actual path and adaptive neuro-fuzzy inference system-generated path is done with respect to x, y, and z directions, respectively. Furthermore, the error between actual and predicted values regarding joint parameters is calculated for a certain arm matrix. The prototype of 5-degrees-of-freedom medical-assisted manipulator is developed at CSIR-CSIO Laboratory Chandigarh, which is also termed as patient-side manipulator to be utilized in robot-assisted surgery. Through the simulation runs, in this work, it is found that the results from adaptive neuro-fuzzy inference system approach are quite satisfactory and acceptable.