基于改进人工势场法的物流无人配送车路径规划  

Path planning for logistics unmanned delivery vehicles based on improved artificial potential field method

作　　者：刘珂 董洪昭 张丽梅 杜秋月 Liu Ke;Dong Hongzhao;Zhang Limei;Du Qiuyue(School of Artificial Intelligence,Beijing Technology&Business University,Beijing 100048,China)

机构地区：[1]北京工商大学人工智能学院,北京100048

出　　处：《计算机应用研究》2022年第11期3287-3291,共5页Application Research of Computers

基　　金：国家自然科学基金资助项目(51805009,52078005)。

摘　　要：传统人工势场法在面对多障环境时易陷入局部最优,导致路径长度和拐点数量增多从而严重降低配送效率。为此,考虑复杂环境多障密集和离散分布特征,结合相对位置检测策略和边界条件判断决策,采用自主虚拟圆作用域设计,修正路径长度和拐点数量缺陷;进一步地,基于模型预测理论,建立四轮转向车辆稳定性控制器,实现理想模型路径的精准跟踪控制;采用MATLAB软件进行模拟避障实验,并与传统算法路径搜索关键要素特征进行对比仿真。研究结果表明,所提出的改进人工势场算法相比于传统算法,全局拐点数量减少10个,路径缩短21.06%,跟踪偏差率低于6%。通过研究,解决了车辆多障环境适配及避障安全性问题,实现了指定轨迹的平稳、安全配送,为无人配送车规划算法的改进提供了理论指导。When confronted with a multi-barrier environment,the traditional artificial potential field method is prone to fall into a local optimum,resulting in an increase in path length and number of inflection points,and thus a significant reduction in distribution efficiency.As a result,this paper combined relative position detection strategy and boundary condition judgment decision,and used autonomous virtual circle field design to correct the defects of path length and number of inflection points in complex environments,taking into account the characteristics of dense and discrete distribution of multiple barriers in complex environments.It also established a four-wheel steering vehicle stability controller based on the model prediction theory in order to achieve accurate tracking control of the ideal model path,and conducted simulated obstacle avoidance tests with MATLAB software,it simulated the key element characteristics of the path search with the traditional algorithm for comparison.In comparison to the traditional algorithm,the proposed improved artificial potential field algorithm reduces the number of global inflection points by 10,the path is shorter by 21.06%,and the tracking deviation rate is less than 6%.The research addresses the issues of vehicle multi-obstacle environment adaptation and obstacle avoidance safety,as well as providing theoretical gui-dance for improving the unmanned delivery vehicle planning algorithm.

关 键 词：无人配送 人工势场法 虚拟圆作用域 路径规划 

分 类 号：TP242.6[自动化与计算机技术—检测技术与自动化装置]