Multi-level path planning algorithm for autonomousrobots

Abstract

The authors present a complete path-planning method for sensor-based navigation in an autonomous mobile robot. The environment is described by probabilistic maps processed with a new obstacle enhancement algorithm. The novelty of the approach consists of using a pyramid structure as a preprocessing module to classical path planning algorithms. This stage enhances the performance of the system, minimising the computational cost by operating in an optimal resolution level. Thus, the system is allowed to work with huge environments with a very fine-coarse descomposition in real time.