Adaptive H-infinity neural network tracking controller for electrically driven manipulators

A new robust learning controller for uncertain rigid-link electrically driv en (RLED) manipulators is presented. This new control scheme integrates H-i nfinity disturbance attenuation design and the direct adaptive neural netwo rks (NN) technique into the well-known computed torque (CT) framework. The role of the NN devices is to adaptively learn the structured and unstructur ed uncertain dynamics. Then, the effects of the approximation error of the NN devices on the tracking performance are attenuated to a prescribed level by the embedded nonlinear H-infinity control. Via a tuning-function-like d esign, each unknown mapping, in the dynamics model of an RLED manipulator, can be learned by only one set NN device in the proposed control structure. For economic reasons, this thrift usage of the NN devices is preferred. Fi nally, a simulation study for a planar two-link RLED manipulator is given. Simulation results indicate that the proposed adaptive H-infinity NN tracki ng controller achieves better tracking performances than the standard CT co ntroller.