Trajectory Regulating Model Reference Adaptive Controller for Robotic Systems

Abstract

The new trajectory regulating model reference adaptive controller (TRMRAC) has been proposed in this brief. The intermediate reference model in the TRMRAC is self-regulated to enhance the stability and robustness of adaptation, with the trajectories of the controlled system proven to be ultimately uniformly bounded. The developed controller is implemented and simulated in a multivariable robotic arm system with its system dynamics approximated by a neural network, showing superior stability characteristics even under unmodeled actuator dynamics and input saturation. To demonstrate its practicality, a nested version of the controller was tested on a quadcopter for quaternion attitude tracking, showing enhanced robustness over the conventional model reference adaptive control strategy.