Research on Fuzzy Adaptive Impedance Control of Lower Extremity Exoskeleton

Qu, Zhicheng; Wei, Wei; Wang, Wei; Zha, Shijia; Li, Tianyi; Gu, Jin; Yue, Chunfeng

doi:10.1109/icma.2019.8816452

Cited by 9 publications

(4 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A key benefit of DRL for exoskeleton control is its ability to learn and provide control in highdimensional continuous joint observations and continuous actuator torque actions. This is achieved by mapping states to actions through deep neural networks approximating the policy and value functions of RL [32]. The use of these continuous observation spaces allows for actuator torque actions to be learned in an end-to-end nature directly from sensory observations of joint parameters [30,56].…”

Section: Deep Reinforcement Learning Control Approachesmentioning

confidence: 99%

“…A simplified dynamics model, however, is not able to represent fully the user-exoskeleton interactions, which can in turn affect the accuracy of the controller [24,53,54]. Therefore, these dynamics models are often unable to precisely model the interactions between the user and exoskeleton [26,[30][31][32][33][34].…”

Section: Introductionmentioning

confidence: 99%

“…Furthermore, the controller must be designed to adapt to different user needs as well as changing needs of the same user during gait rehabilitation. The advantage of model-free DRL is that it does not require a dynamics model and instead only learns through the direct interaction of the exoskeleton with its environment [32]. More information on related works for exoskeleton control is presented in detail in Section 2.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

A model-free deep reinforcement learning approach for control of exoskeleton gait patterns

2021

View full text Add to dashboard Cite

Lower-body exoskeleton control that adapts to users and provides assistance-as-needed can increase user participation and motor learning and allow for more effective gait rehabilitation. Adaptive model-based control methods have previously been developed to consider a user’s interaction with an exoskeleton; however, the predefined dynamics models required are challenging to define accurately, due to the complex dynamics and nonlinearities of the human-exoskeleton interaction. Model-free deep reinforcement learning (DRL) approaches can provide accurate and robust control in robotics applications and have shown potential for lower-body exoskeletons. In this paper, we present a new model-free DRL method for end-to-end learning of desired gait patterns for over-ground gait rehabilitation with an exoskeleton. This control technique is the first to accurately track any gait pattern desired in physiotherapy without requiring a predefined dynamics model and is robust to varying post-stroke individuals’ baseline gait patterns and their interactions and perturbations. Simulated experiments of an exoskeleton paired to a musculoskeletal model show that the DRL method is robust to different post-stroke users and is able to accurately track desired gait pattern trajectories both seen and unseen in training.

show abstract

Section: Deep Reinforcement Learning Control Approachesmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

A model-free deep reinforcement learning approach for control of exoskeleton gait patterns

2021

View full text Add to dashboard Cite

show abstract

“…In [24], a nonlinear disturbance observer was integrated into the backstepping sliding controller to effectively reduce the influence of uncertainties and external disturbances. An impedance control algorithm is proposed in [25][26][27] to realize the stability control of the exoskeleton system. The neural network is another feasible tool to compensate for nonlinear disturbances [28][29][30]; however, prior knowledge of the system is always needed to determine the structure and parameters of the neural network, which makes this method easily affected by dynamic changes.…”

Section: Introductionmentioning

confidence: 99%

Online Adaptive PID Control for a Multi-Joint Lower Extremity Exoskeleton System Using Improved Particle Swarm Optimization

2021

View full text Add to dashboard Cite

Robotic exoskeletons have great potential in the medical rehabilitation and augmentation of human performance in a variety of tasks. Proposing effective and adaptive control strategies is one of the most challenging issues for exoskeleton systems to work interactively with the user in dynamic environments and variable tasks. This research, therefore, aims to advance the state of the art of the exoskeleton adaptive control by integrating the excellent search capability of metaheuristic algorithms with the PID feedback mechanism. Specifically, this paper proposes an online adaptive PID controller for a multi-joint lower extremity exoskeleton system by making use of the particle swarm optimization (PSO) algorithm. Significant improvements, including a ‘leaving and re-searching mechanism’, are introduced into the PSO algorithm for better and faster update of the solution and to prevent premature convergence. In this research, a 9-DOF lower extremity exoskeleton with seven controllable joints is adopted as a test-bench, whose first-principle dynamic model is developed, which includes as many uncertain factors as possible for generality, including human–exoskeleton interactions, environmental forces, and joint unilateral constraint forces. Based upon this, to illustrate the effectiveness of the proposed controller, the human–exoskeleton coupled system is simulated in four characteristic scenarios, in which the following factors are considered: exoskeleton parameter perturbations, human effects, walking terrain switches, and walking speed variations. The results indicate that the proposed controller is superior to the standard PSO algorithm and the conventional PID controller in achieving rapid convergence, suppressing the undesired chattering of PID gains, adaptively adjusting PID coefficients when internal or external disturbances are encountered, and improving tracking accuracy in both position and velocity. We also demonstrate that the proposed controller could be used to switch the working mode of the exoskeleton for either performance or an energy-saving consideration. Overall, aiming at a multi-joint lower extremity exoskeleton system, this research proposes a PSO-based online adaptive PID controller that can be easily implemented in applications. Through rich and practical case studies, the excellent anti-interference capability and environment/task adaptivity of the controller are exemplified.

show abstract