摘要
针对煤矿井下环境的复杂性和不确定性,提出了一种改进遗传算法用于煤矿探测机器人的路径规划。采用栅格法在三维空间中对机器人工作环境进行建模,对染色体编码,初始种群生成、适应度函数的设计等操作进行了改进;算法采用了可变长度的染色体编码方式,使用随机指导式搜索策略来生成初始种群;根据路径长度最短且能耗最少的评价指标设计了适应度函数,并优化设计了遗传算法中的交叉和变异算子,解决了传统遗传算法"早熟现象"和"收敛速度慢"的问题,仿真实验证明了该方法的有效性和可行性。
According to complexity and uncertainty of underground environment,this paper presented an improved genetic algorithm for path planning of coal mine detecting robot.Three-dimensional workspace was modeled by grid method. A series of improvement was made in chromosome coding,population initialization and fitness function design.Variable length coding was adopted,the initial population was generated by the random guided searching strategy.Fitness function was designed with shortest path length and minimum energy consumption as the criterion.Crossover operator and mutation operator were optimized to solve the problems of the simple genetic algorithm such as premature phenomena and slow convergence.The simulation results show that the improved algorithm is effective and feasible.
出处
《太原理工大学学报》
CAS
北大核心
2010年第4期364-367,共4页
Journal of Taiyuan University of Technology
基金
山西省科技攻关资助项目(20080321009)
作者简介
周巍（1969-），女，山西太原人，博士生，讲师，主要从事机电一体化技术及应用研究，（Tel）13593172060
