基于遗传BP算法的温室无线传感器网络定位方法

针对温室移动节点定位的需求,提出了一种基于遗传BP算法的温室无线传感器网络定位方法。该方法主要包括路径损耗指数确定、定位模型训练、未知节点定位3个阶段。首先,锚节点间相互通信,通过高斯校正模型增强定位信息的准确性后,利用最小均方误差估计法确定路径损耗参数;然后,应用遗传BP算法建立未知节点坐标和未知节点至锚节点的距离向量之间的映射关系模型;最后将未知节点接收各锚节点的RSSI值转换为距离向量,输入定位模型中,估算未知节点的位置。试验表明,该方法充分考虑环境对信号传输模型的影响,定位误差≤2m的比例达24%,定位误差≤3.5m的比例达86%,相对定位误差低于4.8%,具有较高的稳定性和定位精度,能够满足实际温室环境的定位需求。

Localization method of wireless sensor network for greenhouse based on genetic BP algorithm

To meet the need of greenhouse mobile node location, a kind of greenhouse wireless sensor network localization method based on genetic BP(back propagation) algorithm was presented. The approach mainly included three stages: path loss index determination, localization model training and unknown nodes positioning. First of all, during the process of anchor nodes communication, path loss parameters could be determined by using minimum mean square error estimation method after applying the gauss correction model to enhance positioning information accuracy; Then, a mapping relationship model between the unknown node coordinate and the distance vector from unknown node to anchor nodes was established by using genetic BP algorith; Finally, the received RSSI value from each anchor node to unknown node could be converted into distance vector, and then entered localization model to estimate the unknown node position. The experiment result showed that the greenhouse wireless sensor network localization method was fully considering the effect of environment on signal transmission model, the localization error rate below 2m was 24%, the localization error rate below 3.5 m was 86% and the relative localization error was less than 4.8%. The method has high ability of stability and precision, and meets the practical need of greenhouse environment localization.