Deep reinforcement learning-based anti-interference control of lower limb exoskeleton
In this paper, we propose a dual-loop control strategy to address the problems of the interference by the human-machine interaction of the lower limb exoskeleton movement. The outer ring adopts admittance control and the human-machine interaction torque is estimated by the generalized momentum observer based on Kalman filter. The inner ring adopts PID control based on DDPG.
In this paper, the simulation platform is built and you can see the simulation results in Fig. 3 to Fig. 9. The wearing experiment of the lower limb exoskeleton is carried out, and you can see the simulation results in Fig. 12 to Fig. 14. The simulation and experimental results show that the designed control strategy makes the lower limb exoskeleton have a good anti-interference effect.