RBFNN与快速非奇异终端滑模的康复外骨骼控制OA

To address the issues of delayed dynamic response caused by prolonged convergence time in control algorithms and in-sufficient trajectory tracking accuracy under external disturbances for lower-limb rehabilitation exoskeleton robots,a Radial Ba-sis Function Neural Network-based Fast Non-singular Terminal Sliding Mode Control(RBFNN-FNTSMC)algorithm is pro-posed to achieve accurate tracking of desired gait trajectories.Nonlinear power terms are introduced into the terminal sliding mode function,enabling the system states to rapidly converge to equilibrium points within finite time.The RBFNN is adopted to approximate and compensate for external disturbances,enabling the exoskeleton to achieve high-precision gait tracking.Based on the Lyapunov stability theory,the stability of the control algorithm is verified.To validate the algorithm's performance,a simulation system integrating the lower-limb exoskeleton dynamics model and control law is constructed on the MATLAB/Simu-link platform,with trajectory tracking control experiments conducted on hip and knee joints.Experimental results show that the proposed RBFNN-FNTSMC algorithm exhibits superior convergence speed and high-precision gait tracking capabilities.Com-pared with traditional sliding mode control,this algorithm effectively eliminates chattering phenomena,improves dynamic re-sponse performance,and enhances system robustness while preserving finite-time convergence characteristics.