室内噪声环境下气味源的多机器人微粒群搜索方法

A PSO-Based Multi-Robot Search Method for Odor Source in Indoor Environment with Noise

针对室内噪声环境下的气味源定位问题,提出一种基于骨干微粒群进化的多机器人协调搜索方法.该方法将每个机器人看作一个微粒,机器人传感器探测到的气味浓度值作为微粒的适应值,所有机器人组成一个进化微粒群;采用动态统计方法在线估计机器人所测气味浓度的噪声强度,并通过区间数表示噪声环境下微粒的适应值;定义微粒间的概率支配关系,更新微粒的全局和局部引导者,并利用关于全局引导者和局部引导者的高斯采样更新机器人的位置.最后,通过2个典型实验环境,验证了所提算法在处理噪声环境下气味源定位问题的优越性.

This paper studies the problem of odor source localization in noise environment,and proposes a cooperative search method of multi-robot based on particle swarm optimization. In this method, a robot is defined as a particle, odor concentration detected by sensors of this robot is regarded as the fitness of this particle, and all robots form the swarm of PSO. Using an improved bare-bones PSO to lead the parlicles search cooperatively odor source, a dynamical statistic method is proposed to estimate noise degree of odor concentration detected by sensors; a probability domination relationship suitable to interval illness is defmed to compare particles and update the local leaders of particles. Moreover, a Gauss sampling method based on the global and local leaders is used to update the positions of particles. Finally, the proposed method is applied to two scenarios with odor sources, and experimental re- suits confirmed its effectiveness on solving the problem of odor SOUlCe localization in noise environment.