基于改进A~*算法的导盲避障路径规划策略研究  

Improved A~* Algorithm for Obstacle Avoidance Path Planning Strategy of the Blind

作　　者：张一豆[1] 赵剡[1] 魏彤[1] 

机构地区：[1]北京航空航天大学仪器科学与光电工程学院,北京100191

出　　处：《航空兵器》2017年第3期86-92,共7页Aero Weaponry

基　　金：国家自然科学基金项目(61233005);国家"973"计划(2014CB744200);北京市科技计划项目(Z15110000211-5022)

摘　　要：路径规划是导盲避障系统的重要组成部分之一。根据导盲避障要求,采用栅格可视化方法与连接代价优先级的设置进行路径规划预处理。在每帧规划中,提出一种改进的动态加权A~*的路径规划算法,通过引入"势场"的概念来动态调节启发函数的权系数,产生新的代价函数来引导路径搜索。采用二次加权的路径转换方法计算当前帧行进方向相对规划路径的角度偏差,作为诱导盲人行走的控制信息。实验表明,改进的动态加权A~*算法能够有效提高路径规划效率,整体路径规划策略可以成功实现导盲避障任务。Path planning is one of the main modules of obstacle avoidance system of the blind. Ac-cording to the request of obstacle avoidance for the blind; the path planning pretreatment is executed by raster visualization and priority setting of connection cost. In each frame of path planning, an improved dynamic weighted A algorithm is put forward. By introducing the concept of potential lieldM to dynami-cally adjust the weight coefficient of heuristic function new cost function is generated to guide the path search. Re-weighting is used for path conversion to compute the angle deviation between the current frame direction and planning path to guide the blind. Experiments show that the improved dynamic weighted A algorithm can effectively increase the efficiency of path planning, and the integral path planning strategy can accomplish the obstacle avoidance task.

关 键 词：导盲避障系统 路径规划 动态加权A算法 势场 代价函数 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]