摘要
针对传统人工势场法在路径规划中存在易陷入局部极小值点、障碍物附近目标点不可达以及固定步长下易发生碰撞的问题,提出一种基于分层监测域的自适应人工势场法(APFLMD)。设计了分层监测域模型,通过建立安全避障范围,实现无人驾驶车辆速度的自适应变化,以提高车辆的避障能力。为避免无人驾驶车辆陷入局部极小值点区域,利用路径点的聚集情况实现局部极小值点的检测。设计了等势圈切线点,并结合二次贝塞尔曲线对路径进行优化,且目标点位于斥力等势圈内部时,提出一种虚拟随机引导策略,以帮助车辆逃离局部极小值点。最后,在斥力场函数中添加距离因子以解决障碍物附近目标点不可达问题。仿真结果表明,在复杂环境下APFLMD算法相较于对比算法,可分别降低49.23%的车辆行驶时间、19.4%的路径长度和车辆能耗以及82.12%的路径平滑度。
To solve the problems of traditional artificial potential field method in path planning,such as easily trapping into local minimum points,unreachable target points near obstacles and prone to collide under fixed step length,an adaptive Artificial Potential Field method based on Layered Monitoring Domain(APFLMD)was proposed.A layered monitoring domain model was designed to realize the adaptive variable speed by establishing the safe obstacle avoidance range,so as to improve the obstacle avoidance ability of vehicle.To avoid the unmanned vehicle trapping into the local minimum point area,the local minimum point detection was realized by using the aggregation of path points.Also,the tangent point of equipotential circle was designed,and the path was optimized by quadratic Bezier curve.When the target point was inside the repulsive equipotential circle,a virtual random guidance strategy was proposed to help the vehicle escape from the local minimum point.The distance factor was added to the repulsion field function to solve the problem of unreachable target points near obstacles.The simulation results showed that compared with the reference algorithm,the APFLMD algorithm in complex environment could reduce the vehicle driving time by 49.23%,the path length by 19.4%,the vehicle energy consumption by 19.4%,and the path smoothness by 82.12%respectively.
作者
潘玉恒
陶艺鑫
鲁维佳
李国燕
王丽
PAN Yuheng;TAO Yixin;LU Weijia;LI Guoyan;WANG Li(School of Computer and Information Engineering,Tianjin Chengjian University,Tianjin 300384,China)
出处
《计算机集成制造系统》
EI
CSCD
北大核心
2024年第5期1908-1918,共11页
Computer Integrated Manufacturing Systems
基金
天津市科技计划资助项目(20YDTPJC00160,21YDTPJC00780)
天津市教委科研计划资助项目(2019KJ101,2017KJ058)。
关键词
无人驾驶车辆
路径规划
人工势场法
监测域
unmanned vehicle
path planning
artificial potential field method
monitoring domain
