Identification and control of nonlinear dynamics of a mobile robot in discrete time using an adaptive technique based on neural PID

FRANCISCO ROSSOMANDO; CARLOS M SORIA

NEURAL COMPUTING AND APPLICATIONS

SPRINGER

Lugar: Berlin; Año: 2014

0941-0643

In this work, original results, concerning theapplication of a discrete-time adaptive PID neural controllerin mobile robots for trajectory tracking control, arereported. In this control strategy, the exact dynamicalmodel of the robot does not need to be known, but a neuralnetwork is used to identify the dynamic model. Toimplement this strategy, two controllers are implementedseparately: a kinematic controller and an adaptive neuralPID controller. The uncertainty and variations in the robotdynamics are compensated by an adaptive neural PIDcontroller. It is efficient and robust in order to achieve agood tracking performance. The stability of the proposedtechnique, based on the discrete-time Lyapunov?s theory, isproven. Finally, experiments on the mobile robot have beendeveloped to show the performance of the proposed technique,including the comparison with a classical PID.