面向校园多机器人协同巡查的路径规划虚拟仿真实验设计

Design of virtual simulation experiment of path planning for multi-robot cooperative patrol in campus

针对校园多机器人协同巡查环境面积广阔、建筑物面积占比大、路径规划“早熟收敛”等问题,该文将A^(*)算法与遗传算法相结合,提出一种多机器人路径规划方法。将校园环境中的多机器人巡查路径规划类比为多旅行商问题,使用混合适应度函数的自适应遗传算法,能较好解决“早熟收敛”问题、提高进化速度和寻找最优解。基于ROS系统,开展了多机器人自主导航的仿真实验,通过构建机器人模型和校园环境模型,设计巡逻方案,实现了在已知路径和环境地图下的多机器人自主导航仿真,验证了算法的可行性和有效性。

Aiming at the problems on multi-robot cooperative patrol,such as large area of campus,large proportion of building area and‘premature convergence’of path planning,a multi-robot path planning method combining A^(*)algorithm and genetic algorithm is proposed in this paper.In this algorithm,the multi-robot patrol path planning in campus environment is regarded as the multiple traveling saleman problem,and an adaptive genetic algorithm with hybrid fitness function is proposed,which can better solve the‘premature convergence’problem,improve the evolution speed and find the optimal solution.Based on ROS system,a multi-robot autonomous navigation simulation experiment is carried out.By constructing the model of robot and environment,and designing patrol scheme,the multi-robot autonomous navigation simulation under known path and environment map is realized,which verifies the feasibility and effectiveness of the algorithm.