Research On Robot Path Planning For Dynamic Environment And Cooperation

Path planning for robots is a key research area in robotics. This paper is based on realistic demand in science projects and takes deep dig into motion planning in dynamic environment and path planning for cooperation of multi-robots system in static environment. The final proposed solutions not only contain traditional artificial potential field and genetic algorithms, but also consider inherent defects in these methods with grid handling and search strategy. Improved or innovative solution is proposed with solid simulation results. Simulation performances prove the validity and robustness of these algorithms. Main contents of this paper are as follows:1. For single robot, an improved artificial potential field algorithm which adapts to collision-free motion planning in dynamic environment is proposed. Besides, some specific details like soft landing issues and local optimum in traditional artificial field method are discussed and taken care of. Simulation results show that the improved artificial potential filed not only reliably gain the path with obstacle avoided but also achieve an optimum path with less waste of robot path length by considering relative velocity and acceleration information between objects. Finally, several scenarios have been created for tests and further validate the robustness of this strategy.2. For multi-robots system, a new off-line cooperative path planning algorithm based on fixed point and genetic algorithm is suggested. This solution takes advantage of parallel computing, feature of local-optimum-free and statistically optimum-path-generation from GA. Furthermore, it combines fixed point method which effectively reduces the wastes of robot path. Simulation results show that the strategy gain more optimum path with less time and generate controlling path for individual robots more easily.3. As the complement of this algorithm, special narrow route is tackled with methods of modeling whole space. Progressive partition of space according to different sub-space is proposed to improve space and time efficiency.