Research And Implementation On Collaborative Motion Planning Algorithm For Multiple Car-like Mobile Robots Which Start Asynchronously

Recently, the research topic about collaborative motion planning for multiple car-like mobile robots in common environment has caught widespread attention, and been applied into variant fields such as traffic system, military, logistics transportation and social activities. Under some specific circumstances, the movement of a car-like mobile robot is driven by the superior scheduling system, which possibly results in variant start time. Besides that, the motion task of each car-like mobile robot may have different degree of urgency. Hence, Comparing with general problem descriptions of this topic, we additionally introduce two new input factors, the start time and passing through priority of each car-like mobile robot. Through reasonable planning laws, we hope to guarantee member car-like mobile robots starting asynchronously and reaching their respective targets independently. Besides that, during the movement of each car-like mobile robot, reducing, and even avoiding, collision with stationary obstacles and other car-like mobile robots is also expected. What’s more, we should likewise make those car-like mobile robots with higher pass-through priorities finish their respective motion tasks as early as possible.In our paper, a central-scheduling, dynamic, realtime algorithm is presented. We decompose the specific motion planing problem into two phases:the trajectory planning one for avoiding collision between car-like mobile robots and stationary obstacles, the trajectory execution one for implementing reciprocal collision avoidance among member car-like mobile robots. Normally in the latter phase, the motion scheme of member car-like mobile robots is moving toward their respective targets totally along with the planned trajectories generated in the former phase. However in our work, as the core content of our work, the second phase would also proceed the Single Scheduling at intervals to configure the latest motion scheme. To be specific, the Single Schduling would calculate the collision results for member car-like mobile robots on condition that they advance according to the current motion scheme and if need be, insert different durations into the motion periods of part of member car-like mobile robots. After the Single Scheduling is completed, member car-like mobile robots would advance according to the latest motion scheme.Simulations and experiments are performed on moderate quanlity (several tens) of car-like mobile robots to demonstrate the effectiveness of the proposed method, which keep consistent with expectations of our algorithm.