改进麻雀搜索算法的移动机器人羽流追踪方法

Plume tracking method of mobile robots based on improved sparrow search algorithm

关键词:  自主羽流追踪  Logistic混沌映射  麻雀搜索算法  Metropolis准则  移动机器人

Key words:  autonomous plume tracking  Logistic chaotic mapping  sparrow search algorithm  Metropolis criterion  mobile robot

摘要:

为改善室内羽流扩散环境下机器人无法获得可靠羽流流向、流速等风信息导致羽流追踪效率低、搜索路径过长的问题,提出一种基于改进麻雀搜索算法的室内机器人自主羽流追踪方法。首先,该方法受麻雀种群捕食与反捕食行为的启发,以羽流质量分数作为个体的适应度值,使寻源机器人在不搭载羽流流向、流速传感器时能高效地追踪羽流并定位源位置。其次,使用Logistic混沌映射分散麻雀种群初始位置,并保留麻雀种群精英解以增加种群多样性；在更新最优解时加入了Metropolis准则提高算法跳出局部极值区域的概率,同时结合改进A*算法优化搜索路径。最后,将改进麻雀搜索算法与遗传算法、鲸鱼优化算法、灰狼优化算法,以及经典麻雀搜索算法进行羽流追踪仿真对比试验,并在物理场景中验证了所提算法的可行性和有效性。结果表明,相比于上述方法,所提方法的成功率分别提升了31.00%、4.84%、1.34%、13.34%,搜索路径长度分别减少了12.098、6.682、4.941、5.448 m。本研究为解决羽流扩散环境下无可靠风信息时移动机器人的高效羽流追踪提供了新的思路和参考。

Abstract:

To address the low plume tracking efficiency and excessively long search paths caused by the inability of robots to obtain reliable wind information, such as plume flow direction and flow velocity, in indoor plume diffusion environments, this paper proposed an autonomous plume tracking method for indoor robots based on an improved sparrow search algorithm (SSA). Firstly, inspired by the predation and anti-predation behavior of the sparrow population, this method used plume concentration as the fitness value of individuals so that the source-seeking robot could efficiently track the plume and locate the source position when it was not equipped with a plume flow direction sensor and a plume flow velocity sensor. Secondly, Logistic chaotic mapping was used to disperse the initial position of the sparrow population, and the elite solution of the sparrow population was retained to increase the population diversity. The Metropolis criterion was added to increase the probability of the algorithm escaping from the local extreme value area when updating the optimal solution. The improved A* algorithm was combined to optimize the search path. Finally, a plume tracking simulation comparison experiment was conducted, in which the improved sparrow search algorithm (ISSA) was compared with the genetic algorithm (GA), whale optimization algorithm (WOA), grey wolf optimizer (GWO), and classic SSA, further validating the feasibility and effectiveness of the proposed algorithm in physical scenarios. The results show that compared with the aforementioned methods, the success rates of the proposed method increase by 31.00%, 4.84%, 1.34%, and 13.34%, respectively, and the search path lengths are shortened by 12.8,6.2,4.941, and 5.448 m, respectively. This study provides a new approach and reference for enabling efficient plume tracking by mobile robots in plume diffusion environments where reliable wind information is unavailable.