Affiliations: School of Mechanical Engineering, Shandong University of Technology, Zibo, Shandong 255000, China
Correspondence:
[*]
Corresponding author: Wenqing Ge, School of Mechanical Engineering, Shandong University of Technology, Zibo, Shandong 255000, China. E-mail: zyj627010@163.com.
Abstract: Aiming at the rehabilitation of lower extremity injury, a control system for the brain controlled medical lower limb exoskeleton was designed. The EEG signal control, pre-programmed control and RBF neural network control method were analyzed. The human-computer interaction control strategy was designed. In order to obtain higher output control accuracy and better robustness, the RBF neural network approximation algorithm with the input signal of both brain control and pre-programmed control was designed. Both the EEG signal recognition experiment and external skeletal prototype control experiment were designed. The results showed that EEG signal recognition accuracy was accurate, and the correct gait movements of external skeleton would be achieved under the active, passive control strategy and the neural network control method. The initial coordination of between human and machine was achieved, which laid a foundation for the further research on stability and rehabilitation training of lower limbs.
Keywords: EEG signal control, lower limb exoskeleton, neural snetworks, control strategy
