李东东, 王雷, 马康康. 基于终距指数的机器人路径规划方法研究[J]. 云南大学学报(自然科学版), 2022, 44(3): 505-514. doi: 10.7540/j.ynu.20210150
引用本文: 李东东, 王雷, 马康康. 基于终距指数的机器人路径规划方法研究[J]. 云南大学学报(自然科学版), 2022, 44(3): 505-514. doi: 10.7540/j.ynu.20210150
LI Dong-dong, WANG Lei, MA Kang-kang. Research on robot path planning method based on terminal index[J]. Journal of Yunnan University: Natural Sciences Edition, 2022, 44(3): 505-514. DOI: 10.7540/j.ynu.20210150
Citation: LI Dong-dong, WANG Lei, MA Kang-kang. Research on robot path planning method based on terminal index[J]. Journal of Yunnan University: Natural Sciences Edition, 2022, 44(3): 505-514. DOI: 10.7540/j.ynu.20210150

基于终距指数的机器人路径规划方法研究

Research on robot path planning method based on terminal index

  • 摘要: 传统蚁群算法在生成信息素浓度时,由于算法生成的路径可能存在冗余成分,信息素浓度可能无法正确反应路径各节点的优劣,蚂蚁无法根据信息素浓度来迅速找出最优路径,导致算法寻优缓慢. 基于传统蚁群算法思想,提出了一种新改进蚁群算法,即通过引入终距指数这一概念,取代信息素浓度的标记功能,蚂蚁可以依赖该指数进行决策选择优良节点. 以20×20的栅格环境地图对改进蚁群算法进行案例仿真,实验结果表明,传统蚁群算法及其他改进蚁群算法分别需要43代及34代才能收敛到最优值,而利用改进蚁群算法仅需要进化3代即可收敛到最优解;为了进一步验证改进蚁群算法的优越性,在对更为复杂的30×30栅格模型仿真,利用传统蚁群算法与其他改进蚁群算法的收敛代数分别为52代与28代,而利用新改进蚁群算法的收敛代数仅为4代;另外,为了进一步验证改进算法的稳定性,对30×30环境模型进行连续运行30次仿真,所需平均收敛代数仅为4.97代.

     

    Abstract: Considering the traditional ant colony algorithm in generating pheromone concentration, because the path generated by the algorithm may have redundant components, the pheromone concentration may not be able to correctly reflect the advantages and disadvantages of each node of the path, so that the ant cannot quickly find the optimal path according to the pheromone concentration, which makes the algorithm search slowly. Based on the idea of traditional ant colony algorithm, a new improved ant colony algorithm is proposed, that is, by introducing the concept of terminal distance index to replace the marking function of pheromone concentration, so that ants can rely on the index to make decisions and select good nodes. A 20×20 grid environment map is used to simulate the new improved ant colony algorithm. The experimental results show that the traditional ant colony algorithm and other improved ant colony algorithms need 43 generations and 34 generations to converge to the optimal value respectively, while the new improved ant colony algorithm only needs 3 generations to converge to the optimal solution, which proves the advantage of the algorithm in the speed of optimization. In order to further verify the superiority of this algorithm, for the more complex 30×30 grid model, the simulation results show that the convergence algebra of the traditional ant colony algorithm and other improved ant colony algorithm need 52 and 28 generations to converge to the optimal value respectively, while the new improved ant colony algorithm only needs 4 generations to converge to the optimal solution. In addition, in order to further verify the stability of the improved algorithm, the 30×30 environment model is simulated for 30 times, and the average convergence algebra is only 4.97 generations.

     

/

返回文章
返回