Anti-disturbance control of CPG bionic reflection in pneumatic muscle actuator

Abstract

Addressing the joint control problem of pneumatic muscle-driven robots, this study aims to design a bionic reflex mechanism to enhance the robots’ adaptive capacity to various disturbances. Based on the biological reflex mechanism, we developed a spindle reflex and deep tendon reflex control system based on CPG (central pattern generator) to mitigate the sudden impact on the hip joint and the continuous blocking force on the knee joint, respectively. The spindle reflex controller incorporates the fast response of sliding mode control to effectively minimize the trajectory deviation of the hip joint under impact disturbances. The deep tendon reflex controller integrates RBF neural network adaptive control and the Tegotae framework to suppress excessive tension in the knee joint and augment the system’s adaptability to the blocking force disturbances. The experimental results confirm that the two reflex mechanisms significantly enhance the robustness and flexibility of the pneumatic muscle-driven robot in motion.