Dynamic Algorithm Selection for Mobile Robots Motion Planning

Abstract

Selecting the appropriate algorithms to construct a synthetic planner for mobile robots in a given environment is a very beneficial task. The planner of mobile robots is a highlyconfigurable system, which mainly includes obstacle avoidance and navigation algorithms. This paper presents a framework to dynamically combine the synthetic planner according to the robot surrounding environment and the robotic states. Furthermore, two feasible approaches are introduced to select, i.e., the blackbox approach and the rule-based approach.