An Extension of the Distance-Propagating Dynamic System for Robot Path Planning to Safe Obstacle Clearance

Abstract

In this paper we extend our previously presented efficient distance-propagating dynamic system for real-time robot path planning in dynamic environments to the case where safety margins around obstacles are included. Inclusion of safety margins approximately triples the number of arithmetic operations, however, the distance-propagating dynamic system is still very computationally efficient. The algorithm uses a grid representation of the environment, which need not be regular, and is applicable to dynamic environments where both targets and obstacles are permitted to move. No prior knowledge of target or obstacle movement is assumed. Safety margins around obstacles are implemented as "soft" margins defined by local penalty functions around obstacles which represent the extra distance the robot is willing to travel in order to avoid passing through this margin. The path through which the robot travels minimizes the sum of the current known distance to a target and the cumulative local penalty functions along the path. The effectiveness of the algorithm is demonstrated through a number of simulations.