A discrete hybrid nonlinear-model referenced adaptive controller for robotic manipulators

A hybrid controller for robotic manipulators was designed and tested using a distributed multiprocessor architecture and a PUMA 550 robot arm. The feedback control signal consists of two components. The first component is computed using the nonlinear dynamic equations and is thus based on a computed torque approach. Due to modeling errors in the dynamic equations, however, a second component of the control signal is necessary to compensate for these errors. This second component is derived using a model referenced adaptive approach. The manipulator system is considered to be composed of a nominal part characterized by the dynamic equations used to compute a nominal input value and a variational part describing the error between the behavior of the nominal model and the actual system behavior. A model referenced adaptive feedback control signal is derived from the observed output error and superimposed on the nominal input as a variational term of the input.<
>