Ensuring path tracking stability of mobile robots in harsh conditions: An adaptive and predictive velocity control

The aim of a mobile robot path tracking algorithm is to ensure that the desired path is followed as accurately as possible. This problem has been intensively studied in literature with satisfactory results in on-road context. Nevertheless, performances may be depreciated when the expected ideal conditions are no longer satisfied, as it is the case when moving off-road: in such a context, bad grip conditions together with actuator saturations may generate significant perturbations, especially at high speed. Beyond a lack of accuracy, instabilities (such as spin around or non-controllability) may arise. This paper proposes an adaptive and predictive approach in order to preserve the path tracking stability thanks to the modulation of the robot velocity. Relying on the on-line observation of the grip conditions and the reference path properties, the maximal velocity admissible in a near future is computed and applied, if necessary, instead of the desired speed. A steering angle control law, designed to be independent of the robot speed, acts in parallel. The capabilities of this algorithm are tested through actual experiments with a mobile off-road platform.