摘要
针对传统全向移动机器人局部路径规划算法无法提前对行人轨迹进行预测而导致智能性差的问题,提出了基于行人轨迹预测的全向移动机器人路径规划算法。在使用激光雷达实时对行人进行识别定位的基础上,利用卡尔曼滤波算法对行人下一刻全局坐标进行预测,并根据预测结果调整机器人局部路径规划,使其尽快逃离行人预测坐标点辐射圈,同时避开行人行走路径,在提前响应动态障碍物的基础上避免了二次避障和无局部最优解而导致的急停。最后进行行人与机器人交叉而行的综合实验,实验对比曲线证明建立的算法可行且智能性较高。
ABSTRACT:Aiming at the problem that the traditional path planning algorithm of traditional omni-mobile robot can not predict the pedestrian trajectory in advance, we put forward a path planning algorithm of 0mni-mobile robot based on pedestrian trajectory prediction. Based on the real-time recognition and localization of pedestrians using the lidar, the Kalman filter algorithm was used to predict the global coordinates of pedestrians at the next moment, and the local path planning of the robot was adjusted according to the prediction results, so as to escape the pedestrian projections and avoid pedestrian walking path. Thus, the emergency stop caused by the second obstacle avoidance and the local optimal solution can be avoided on the basis of avoiding the dynamic obstacle in advance. Finally, a comprehensive experiment was carried out for pedestrian and robot crossing. Experimental comparison curves show that the proposed algorithm is feasible and intelligent.
出处
《计算机仿真》
北大核心
2018年第1期310-314,共5页
Computer Simulation
基金
国家自然科学基金(51005126)
关键词
全向移动机器人
路径规划
行人轨迹预测
避障
Omni-mobile robots
Path planning
Pedestrian trajectory prediction
Obstacle avoidance
作者简介
刘召(1979-),男(汉族),天津人,博士,主要研究领域为机器人与仿生学,特种机器人,机电自动化及系统集成;宋立滨(1978-),男(汉族),山东省潍坊市人,博士,助理研究员,主要研究领域为特种机器人技术与应用、生物医疗仪器;于涛(1989-),女(汉族),河北省承德市人,硕士,主要研究领域为智能化自动化系统与装置;耿美晓(1990-),女(汉族),河北省衡水市人,硕士,主要研究领域为计算机技术。