摘要
移动机器人的工作环境复杂且多变,决定了路径规划在移动机器人研究中的重要地位。传统的求解方法均无法同时优化多个目标且各自存在缺点,而SPEA2算法则非常适合求解存在多个优化目标的机器人路径规划问题,具有参数少、解集分布均匀的优点。但同时也存在早熟收敛和收敛速度慢的问题,从而影响了路径规划效率。针对上述缺陷,对SPEA2算法加以改进,提出了采用种群多样性的自适应遗传概率调整公式,并且加入修复和平滑算子以提高路径规划效果。仿真结果表明,改进算法相比于经典SPEA2在收敛能力上有了较大的提高,得到的机器人行走路径也非常理想,为机器人路径规划的优化提供了参考。
The robot path planning plays an important role in mobile robotics research. Because of the complicated working environment, the current main solutions cannot optimize multiple goals simultaneously and their own shortcomings exist. The SPEA2 is well suited for solving the mobile robot path planning problems with multiple goals, which has the advantages of less parameters and even distribution solution sets. But at the same time, there are also premature convergence and slow convergence defects affecting the efficiency of path planning. To solve this problem, an improved adaptive SPEA2 algorithm was proposed in the paper, which involves an adaptive adjustment strategy which changes genetic parameters based on diversity of population, and special operators for robot path planning. The simulation results show that the improved SPEA2 algorithm convergence ability has been greatly improved, and all the robots walking paths are ideal.
出处
《计算机仿真》
CSCD
北大核心
2014年第7期346-350,共5页
Computer Simulation
基金
国家自然科学基金资助项目(61105115)
