林中无线有向感知传感器网络信号传输机理及节点布设方法

Forest ecological monitoring which based on directed perception wireless sensor networks is currently a forefront research topic in forestry ecology field. This research includes multiple-subject intersectional field such as ecological environment, information perception, and so on. Node deployment problem is the key to promote the application of directed perception wireless sensor networks in forestry filed. This project aims at a scientific issue: radio frequency signal transmission attenuation mechanism of directed perception wireless sensor networks, we via theoretical analysis and experimental research, try to determine the main factors which affect 2.4GHz radio frequency signal transmission in the forest. Based on wavelet algorithm data regression, the 2.4GHz wireless radio frequency signal propagation loss models for the main influence factors of a single variable will be established respectively. Based on the neural network theory, a composite model will be established to estimate the node's maximum effective communication distance. The connected coverage problem is simplified assuring perceived sectors to cover a given area, and ensuring the mutual communication among sectors. Considering the influences of node communication radius, sensing radius and perception angle to connected coverage scheme, we via theoretical derivation and simulation platform, confirm the node optimal deployment mode which is oriented to forestry application. By these researches, a theoretical innovation and technological breakthroughs for radio frequency signal transmission mechanism and node layout methods of directed perception wireless sensor networks in the forest can be achieved, the application of wireless sensor networks in the field of forest ecological monitoring should be promoted.

基于无线有向感知传感器网络的森林生态监测是当前国际前沿研究课题，涉及林业生态、信息感知等多学科领域交叉。节点部署问题是无线有向感知传感器网络在林业领域推广应用的关键问题。本项目针对无线有向感知传感器网络射频信号传输衰减机理这一科学问题，通过电磁波理论分析与试验研究，确定林中影响2.4GHz射频信号传播的主要因素；基于小波算法数据回归分别建立2.4GHz无线射频信号针对主要影响因素的传播损耗单一变量模型；基于神经网络理论建立复合模型，估算节点间最大有效通信距离；将连通覆盖问题简化为用感知扇形覆盖给定区域，并要保证扇形间可相互通信；兼顾考虑节点通信半径与感知半径之比及感知角度对覆盖连通方案的影响，通过理论推导与专业仿真平台，确定面向林业的节点最优部署方式。本项目可实现无线有向感知传感器网络射频信号在林中传输机理及节点布设方法方面的理论创新与技术突破，促进无线传感器网络技术在森林生态监测的应用。

基于无线有向感知传感器网络的森林生态监测是当前国际前沿研究课题，涉及林业生态、信息感知等多学科领域交叉。无线通信传输信号传播特点、规律和节点部署问题是无线有向感知传感器网络在林业领域推广应用的关键问题。.针对无线图像传感器网络在林区的无线信号衰减规律，项目组通过研究电磁波理论分析与试验研究不同林相、立地条件、气候因素等条件下的2.4GHz 无线有向感知传感器网络射频信号传输衰减规律，建立了相关小波回归分析、BP神经网络、SVM等信号场强预测模型，估算节点最大有效通信距离。经实际林区的测试及模型估计的预测，本项目平均预测最大有效通讯距离为63.99米。项目组对无线传感器网络中有向感知网络的部署方式展开研究，建立了适合于林中应用的有向感知节点模型，同时将连通覆盖问题简化为用感知扇形来覆盖给定区域，在保证扇形之间可以相互通信的同时，兼顾考虑有向感知节点通信半径与感知半径之比以及感知角度对覆盖连通方案的影响。项目组在保证连通覆盖的前提下，研究了正三角形、正四边形、菱形、正六边形及非正六边形等部署方式，保证有效监测覆盖的同时，使监测节点数目最少。项目组采用部署成本和有效覆盖的判别方法,确定平铺部署和圆形部署两种方式，发现当感知夹角大于1.073rad(即61.5°),圆形部署方式冗余覆盖面积小于平铺部署,在圆形部署方式研究中采用规格算法,当通信半径小于感知半径,采用正六边形部署方式,使用节点数目最少,冗余覆盖面积最小。项目组对概率感知模型在林区应用中的问题进行了分析,确定监测区域的概率阈值,依次改进优化虚拟力算法、粒子群算法和混沌粒子群算法,对部署进巧调整,使得监测区域无感知盲区,且提高感知概率分别达到82.7%、79.23%、85.61%。.本项目完成了原定的技术指标，实现了无线有向感知传感器网络射频信号在林中传输机理及节点布设方法方面的理论创新和技术突破，为无线传感器网络技术在森林生态监测的应用提供了理论基础。