Direct model reference adaptive control of a PUMA manipulator

摘要

The results of controlling a PUMA 560 robotic manipulator using a Command Generator Tracker (CGT) based model reference adaptive controller (DMRAC) are presented. Initially, the DMRAC algorithm is run in a simulation using a detailed dynamic model of the PUMA 560. The algorithm is tuned on the simulation and then used to control the manipulator using minimum jerk trajectories as the desired reference inputs. Performance achieved in both simulation and on the actual robot shows that the algorithm is robust in the presence of sudden load changes.