Path planning with uncertainty for car-like robots

Abstract

This paper presents the first path planner taking into account both nonholonomic and uncertainty constraints. The case of a car-like robot subject to cumulative and unbounded configuration uncertainty related to bounded control and sensing errors is considered. Assuming the existence of landmarks allowing the robot to relocalize itself in particular places, we present an algorithm that computes paths that are both feasible and robust. In other words, they respect the nonholonomic constraints of a car-like robot and the robot is assured to reach its goal by following them, as long as its control and sensing errors remain within bounded sets.