摘要
通过在弧长-曲率空间建立车辆运动学模型的方法,在满足非完整约束条件的基础上,将运动规划问题转化为函数优化问题。为提高PSO算法的优化速度,满足算法工程应用的实时性要求,提出一种基于多任务种群协同进化的粒子群优化算法。该算法将种群分为3种执行不同任务动作的子群,充分扩展搜索范围,挖掘搜索域内的有用信息,使种群的全局搜索能力和局部搜索能力达到较好的平衡状态。实验结果证明,将协同进化PSO算法应用于弧长-曲率空间中的函数优化问题,可实现对自主车辆的运动规划,规划轨迹满足车辆运动学和动力学约束,保证了车辆行驶的安全性和平稳性。
The motion planning for the autonomous land vehicle in an environment including obstacles is a nonholonomic constraint global optimization problem. For this optimization problem, the motion model of the vehicle is built firstly in the arclength-curvature space, and then the motion planning problem is transformed to a function optimization problem in this paper. To improve the optimiza- tion speed of PSO algorithm and satisfy its real-time requirement in engineering application, an improved PSO based on the multi-tasking subpopulation cooperation is developed. By introducing the idea of multi-tasking subpopulation mechanism used in artificial bee colony algorithm, the population is divided into three subpopulation with different task, and the search area is extended and more useful infor- mation is excavated without the increase of the amount of the particles. The improved PSO-MTC algorithm is used to optimize the param- eters of arclength-curvature model. The simulation result shows the effectiveness of the proposed method in solving the problem of ALV motion planning.
出处
《控制工程》
CSCD
北大核心
2013年第1期169-174,共6页
Control Engineering of China
基金
军队战略投资科研项目
关键词
自主车辆
运动规划
弧长-曲率空间
粒子群优化
非完整约束
协同进化
autonomous land vehicle
motion planning
arclength-curvature space
particle swarm optimization
nonholonomic con-straint
cooperative evolution
