基于栅格法的矿难搜索机器人全局路径规划与局部避障

Global path planning and local obstacle avoidance of searching robot in mine disasters based on grid method

针对矿难发生后井下环境的不确定性,提出一种以矿难前的GIS(Geographic information system)地图为基础建立环境栅格模型并结合改进遗传算法的矿难搜索机器人全局路径规划方法。效仿蚁群算法中的信息素提出基于位置信息负反馈的方法,并结合优先权分组的思想,提出一种新的有效的种群初始化方法,同时将该种群初始化方法应用到变异算子中,且依据最优解的变化情况自适应地调整交叉和变异的概率。与此同时,针对环境信息的不同变化情况,结合全局路径规划结果对机器人进行局部避障方法的研究。最后,通过仿真实验证明本方法能够快速有效地在已知环境中得到机器人的最优路径,并且能够在局部变化的环境中实现实时避障。

Aiming at the uncertainty of the environment in mine disasters,the gird model was built based on the GIS（geographic information system） map acquired from the mine in advance,and a modified genetic algorithm was provided for global path planning.An efficient method for population initialization which adopted the position information negative feedback like the ant colony optimization and priority grouping was provided.Also the population initialization method was applied to the mutation operator.The method self-adaptively adjusted the probabilities of crossover and mutation due to the change of the best resolution.According to the condition of the environment and combining with the global path planning result a local obstacle avoidance method was presented.Finally,the simulation results verify that the provided method can provide an optimal path in known environment effectively,and reallize the real time obstacle avoidance in locally changed environment.