Cooperative Path Management for Mobile Systems Based on Adaptive Dynamic Programming

Abstract

Path-planning for multiple vehicles is a topic relevant to many areas of research including navigation and robotics. In this paper, we present a novel approach for cooperative movement management to examine how two objects can orchestrate their movements so as to avoid collisions and retain a good chance of returning to their intended paths. When the objects recognize that they are at risk of a collision, they cooperatively change course to avoid hitting each other and return to their original course when the risk is averted. The paths are generated and selected by the system after learning the approximate reward of each movement composition, which takes into account the smoothness of paths, as well as the distances between, and velocities of, the vehicles. The path-planning problem in this research is treated as a state transition process in a continuous high-dimensional system. Adaptive (or Approximate) Dynamic Programming (ADP) is applied to solve path-planning task. In ADP, an approximate value function for the entire search space defined for the system is heuristically developed according to certain rules. The method of ordinary least squares serves as the regression mechanism to approximate the value function for the entire search space, providing a scoring technique for the discrete system featuring finite states. This paper summarizes the concept and methodologies used to implement an online cooperative collision avoidance system. Different scenarios are tested to assess the performance of the proposed algorithm.