摘要
由于救援机器人处于复杂环境中,面向紧急救援任务的实时性要求较高,因此,救援机器人的路径规划技术在整个救援过程中发挥着十分重要的作用.针对复杂环境条件,用多边形表示障碍物,设计了一种基于障碍物编码的遗传算法,进行路径规划.与以往的基于顶点编码的方法相比,该方法对障碍物复杂不规则的情况适应性更强,同时也缩减了解搜索空间,提高了算法的效率,增强了路径规划的实时性.通过异构多机器人联合救援模拟实验验证和分析,表明所提出的方法能够使机器人避让复杂环境中不同类型的障碍物,实现高效实时的路径规划,可操作性强,可以推广应用于实际救援系统中.
Rescue robots often work in complex environments and their effectiveness can determine whether victims will survive. This makes research on robot path planning of great significance. In this paper, obstacles in complex environments were modeled as polygons. Based on the encoded polygon, a new path planning algorithm was proposed using genetic algorithms. Compared with previous vertex-based encoding methods, the proposed encoding method can more easily adapt to the irregular shapes of obstacles. The solution space is thus considerably reduced, improving the efficiency and real-time performance of the path planning algorithm. Experimental results demonstrated that the algorithm can effectively guide a robot around obstacles in complex environments, producing acceptable paths for robots covering difficult terrain in rescues. The algorithm can be extended to practical rescue systems involving multiple robots.
出处
《智能系统学报》
2009年第5期414-420,共7页
CAAI Transactions on Intelligent Systems
基金
国家自然科学基金资助项目(60875072)
国家高技术研究发展计划(863)资助项目(2006AA04Z207)
教育部博士点基金资助项目(20060006018)
中澳国际合作资助项目(2007DFA11530)