Finite-Time Command Filtered Adaptive Control for Robot Manipulators in Random Vibration Environment

Abstract

This brief introduces a finite-time tracking control algorithm for robot manipulator systems in a random vibration environment, which addresses the challenges of parameter uncertainty and input saturation. The algorithm combines command filtered adaptive backstepping with neural networks to approximate unknown nonlinear dynamics and avoid the singularity problem of traditional finite-time backstepping methods. An error compensation mechanism based on the fractional power function is also introduced to improve trajectory tracking accuracy, and the algorithm is shown to ensure practical finite-time stability in mean square. Numerical simulations demonstrate that the effectiveness of proposed method.