Robotic manipulators decoupling control based on ANN a th-order inverse system method

A composite control algorithm based on ANN (artificial neural networks) a th-order inversion and PID control strategy was proposed for the trajectory tracking of robotic manipulators with unknown dynamics. In the scheme, the ANN a th-order inversion was used to approximately decouple the controlled nonlinear robotic manipulators system into a number of independent SISO (single input single output) linear subsystems. Neighborhood-based Levenberg-Marquardt algorithm was used for ANN training. Then the PID algorithm was used to compensate for the system errors, thus the system stability is guaranteed and its tracking errors are weakened. Simulation results show that the scheme not only improves the tracking performance, but also has a certain self-adapting ability.