Adaptive nonlinear model following and avoidance under uncertainty

A nonlinearizable plant (several equilibria, strong nonlinearity, nonlinearly coupled) with uncertain but bounded parameters is controlled adaptively to follow or/and avoid a compatible reference model, in real (possibly stipulated) time. The technique used is that of the model reference adaptive control (MRAC) extended to the case at hand. The extension abandons the classical error equation. Instead it secures real-time convergence to the diagonal in the Cartesian product of the plant and model parameter-state spaces. Algorithms for signal adaptive feedback controller and adaptive laws are given. The results are applied to a robotic RP-manipulator with uncertain payload and inertia