Self-Tuning Subordinate Control System Based on Neural Network for Nonlinear Electric Drivers of Lower Limbs of Exoskeleton

Authors:

Abstract
This article proposes a subordinate system control with a self-turning adaptation of the coefficients of the PID controller of position control loop by an analog neural network for nonlinear electric drive of lower limb exoskeleton (LLE) under the influence of uncertain parameters of the dynamic model, nonlinear elements (friction, elasticity of the dynamic model), and stochastic external disturbances. To correct the coefficients of traditional PID controllers, it is proposed to use a neural network (NN), which allows self-tuning the coefficients directly from its outputs based on errors of the control system with the possibility of self-determination of the Jacobian matrix of the neural network. The paper also produces the dynamics model of exoskeleton movement with five links of two legs and a body, considering nonlinear electric drives and external disturbances acting. The efficiency of the proposed control method with the joint motion angular response of exoskeleton is proved by results simulation in the environment Matlab/ Simulink.