Tracking control of industrial manipulator based on adaptive RBF neural network with local model approximation

A Lyapunov-stability-guaranteed local model adaptive RBF neural network control algorithm is proposed for industrial robots. This algorithm eliminates the need for an exact plant mathematical model, enabling real-time learning and compensation of system nonlinearities and uncertainties. An adaptive control law adjusts neural network parameters online to achieve high-precision tracking control of robot manipulators. Particle Swarm Optimization (PSO) is employed to optimize the RBF basis width parameters. Validation was performed using the ABB IRB1600 industrial robot, modeled and simulated in MATLAB Simscape. Real-time trajectory tracking was demonstrated, with ADAMS co-simulation used for further verification. Results demonstrate the algorithm's effectiveness in reducing tracking errors and enhancing robustness and adaptability, while maintaining stability within the Lyapunov framework.