摘要
核电站一般建在偏远的地方,地势开阔,可供人通行的路径多,实时监测困难。现有核电站实物保护系统的监视范围较小,对核电站的整个周界不能形成有效地实时监视。这就让不受地域和环境限制的无线传感器网络大有可为。但由于节点随机撒布和复杂多变的环境干扰,使得节点上传感器的检测概率和虚警概率难以先验获取,从而导致基于传感器特征已知的检测算法在无线传感器网络应用中面临困难。本文提出了一种针对动目标检测的方法——区域投票法。该方法通过搜索特殊三角形外心,确定目标经过区域,而后通过在特定圆形区域内使用投票法来检测目标是否存在。仿真结果表明,与全区域节点加权法相比,在复杂环境下,所提出的目标检测方法可以达到高检测概率,同时可以有效降低虚警率。完全满足核电站对周界实时监测的需要。
The nuclear power plant （NPP） is commonly constructed in distant suburbs which has wide surrounding space. There are many routes to approach the NPP so that is very difficult to take surveillance to all of routes. The existing physical protection system can only monitor a small region. All of these give good opportunity to wireless sensor networks （WSN） which is not limited to region and surrounding. But, the random distribution of the nodes and the intricate disturbance of the environment makes difficult to obtain the prior of the detection and false alarm probability of sensors. Accordingly, difficulty is brought into the detection algorithms that based on the characteristics of sensors. A new target detection method -- Local Voting -- is proposed in this paper. It first finds out the object region by seeking the centroid of a peculiar triangle. Then, the object is determined based on a voting method in a given circular region. Simulation results show that the proposed local voting method can effectively has high detection probability and low false alarm rate in intricate environment compared with the global weighting method. It can fully satisfy the demand of real time surveillance on perimeter around NPR
出处
《自动化博览》
2015年第9期74-77,共4页
Automation Panorama1
关键词
核电站
实物保护系统
无线传感器网络
特殊三角形
动目标
区域投票法
Nuclear power plant
Physical protection system
Wireless sensornetworks
Peculiar triangle
Moving target
Local voting method
