Non-Linear Prediction-Based Trajectory Tracking for Non-Holonomic Mobile Robots

This work addresses the trajectory tracking problem for a non-holonomic differential drive mobile robot with a constant time delay
$h$
at the input signal. To compensate for the adverse effects of the input time delay on the vehicle, a non-linear prediction-observer scheme based on a sub-prediction strategy that asymptotically estimates the future values of the state,
$h$
units of time ahead was introduced, and, thanks to the characteristics of the system, a condition which depends only on the gains of the predictor-observer is obtained for the convergence of the predicted states. Non-linear feedback based on the estimated future state is proposed to tackle the trajectory tracking problem of a mobile robot. The closed-loop system describing the prediction strategy and trajectory tracking solution was formally analyzed, showing the asymptotic convergence of the prediction and tracking errors to the origin. Numerical and real-time experiments were performed to evaluate the prediction-based control scheme, which show adequate performance.