基于移动传感器网络的气体源定位

Gas Source Localization Based on Mobile Sensor Network

针对目前基于无线传感器网络定位气体源方法中存在的因探索区域尺寸、边界等先验知识缺失而造成节点部署困难的问题,提出了一种使用有限个低成本机器人构成移动传感器网络进行气体源定位的方法.该方法中,首先以期望拓扑结构排列的移动传感器网络节点采集气体浓度并上传至上位机;然后上位机使用非线性最小二乘方法对气体源位置进行估计,并将位置估计结果发送给移动传感器网络节点;最后由移动节点计算出目标位姿,并采用饱和控制的方法镇定到该位姿.通过上述步骤的不断循环,移动传感器网络最终将移动到气体源附近,从而实现对气体源较为精确的定位.室内自然通风环境中,使用6个自制的机器人组成具有固定圆形拓扑的移动传感器网络进行了气体源定位实验.实验结果表明,适当选择拓扑半径,该方法可在6,min左右实现精度为30,cm的气体源定位.

Considering the difficulty in deploying the nodes for the wireless-sensor-network-based gas source local-ization（GSL）methods owing to the lack of the prior knowledge such as the detecting area and the boundary,this paper addresses a GSL method using the mobile sensor network（MSN）which is composed of a limited number of low-cost robots. In this method,firstly,the MSN nodes with expected topology collect gas concentration and send it to a central node;Secondly,the central node estimates the gas source location using a nonlinear least square method and sends the outcome to the MSN nodes;Finaly,the MSN nodes calculate the target pose and the nodes are stabi-lized near the target using saturation control. Through repeating the above cycles,the MSN nodes can eventually approach the gas source. Experiments in an indoor ventilated environment show that a GSL accuracy of 30 cm within about six minutes could be achieved.