Planification de mouvements avec prise en compte explicite des incertitudes géométriques

Abstract

A new motion planning approach is proposed for a holonomic robot (with 2 degrees of freedom in translation and 1 in rotation) under geometric uncertainty constraints. The constraints take into account the position, orientation and motion of the robot. The original contribution of the approach is the resolution method and the modelling of uncertainties. The approach consists of two functions defined in the configuration space: (1) an artificial potential function used to generate incremental motions in both the free space and the contact space; (2) an exploration function that analyses the contact space in order to overcome the local minima of the potential function. ``Contact'' between the robot and the obstacles is used to reduce the position and orientation uncertainty of the robot. An explicit uncertainty analysis using geometric models is integrated into the approach. The approach is validated with two robotic tasks: (1) the motion planning for a free flying object on a plane, and (2) the navigation of a mobile robot.