摘要
针对已有机器人全局路径规划方法中存在的知识利用不充分问题,借鉴文化算法的双层进化结构,提出一种融合进化知识和角度信息的新型全局路径规划方法。根据问题需求,提出常识知识、角度信息和进化知识三类知识描述。根据各类知识特性不同,分别用于约束环境、指导个体可行性判断和修复算子。针对具有不同障碍物的两类环境,通过仿真分析与对比,表明本方法可以有效降低不可行个体判断和修复中的计算复杂度,提高进化收敛速度和解性能。
Existing global path planning methods do not utilize knowledge enough.To solve this problem,a novel path planning method based on evolution knowledge and angle information was proposed by adopting dual evolution structure in culture algorithms.Considering the need of the problem,three kinds of knowledge,including common sense,angle information and evolution knowledge,were given.These knowledges were used to constrain environment,judge and repair infeasible individuals according to their characters.Taking two types of environments with different obstacles as examples,simulation results indicate that the algorithm can decrease the computation complexity for judgment and repair of infeasible individual.It also can effectively improve the speed of convergence and have better computation stability.
出处
《系统仿真学报》
CAS
CSCD
北大核心
2010年第5期1142-1147,共6页
Journal of System Simulation
基金
国家自然科学基金(60805025)
国家863计划项目(2007AA12Z162)
中国博士后科学基金项目(2005037225)
关键词
路径规划
角度信息
进化知识
遗传算法
栅格法
path planning
angle information
evolution knowledge
genetic algorithm
grid model
