作　　者： (贾春雪）; (罗琦）; (龚杨杨）;

机构地区： 南京信息工程大学信息与控制学院,南京210044

出　　处： 《计算机科学》 2017年第9期290-295,共6页

摘　　要： 传统的多智能体避障算法在考虑障碍物形状时存在路径冗余、能耗高等现象,不具备普适性。为此,首先采用自动识别凸形化的方式对不规则障碍物进行形状规则化的转变;其次,受子目标思想的启发,将智能体行走的路径转化成规则化后障碍物的多个登陆点路径的叠加,从而保证每段路径的最优化,再选取全局最优路径;最后,利用MATLAB进行仿真,对比与分析了另外两种算法执行的结果,验证了算法的可行性和有效性。 The phenomenon of path redundancy and high energy consumption exist in the traditional multi-agent obsta cle avoidance algorithms when the shape of the obstacle is considered,and the algorithms are not universal. Therefore, firstly,the method of automatic recognition convexity was defined to transform the obstacle from irregular to rule. Se condly,inspired by the idea of sub-target,the path of the agent was transformed into the superposition of multiple lan ding points of the obstacle after being ruled,so as to ensure the optimization of each path,and then selected the global optimal path. Finally, MATI.AB was used to simulate, compare and analyze the results of the other two algorithms,and the feasibility and effectiveness of the algorithm was verified.

关 键 词： 不规则 路径规划 自动识别凸形化 登陆点