基于无线磁阻传感器网络的车辆检测技术研究

发布时间：2018-09-18 15:09

【摘要】：随着我国城市化的加快，城市中机动车保有量的持续高速增长，交通问题日益严峻。国内大中城市的停车调查数据表明，城市路边停车问题尤其突出，特别是繁忙时段，动态交通拥堵严重。在此背景下，智能交通系统（ITS）逐步形成。在智能交通系统中，交通信息的采集（如车辆检测）占有重要地位，是交通流预测、控制及紧急事件快速反应的基础。由于车辆是铁磁性物质，车辆在空间的存在会对地磁场产生扰动，故而可以通过装置在地面上的各向异性磁阻（AMR）传感器探测该扰动，从而达到车辆检测的目的。将AMR传感器节点与无线网络相连，构成无线传感器网络（WSN），可以广泛应用于智能交通系统。目前，基于无线磁阻传感器网络的车辆检测技术还不成熟，检测车辆泊车等基础性问题依然没有得到很好的解决。本文针对上述问题开展研究工作，，首先通过对车辆泊车对地磁场扰动信号的提取及考察，分析信号的特征及其识别方法；其次通过提取的信号特征数据，利用协同信息处理策略，融合相邻节点数据，在现有的车辆检测算法的基础上，针对单节点检测提出局部极值检测算法（REA），针对多节点检测提出协同决策检测算法（CDA）。局部极值检测算法采用了基于过程的方法，考虑泊车信号变化过程的情况，通过状态机实时提取信号的波动特征，根据特征数据设计相应的判断规则。协同决策检测算法关注相邻停车位信号的关联性，在传感器节点检测到的信号幅度较小时触发检测，根据REA算法得到的特征数据，通过路由器融合不同节点的信息做出综合判断。上述两种算法分别应用于传感器节点和路由器节点。算法在实际的系统中应用超过6个月，节点数超过100个，通过实验验证和实际系统的反馈数据，证明了所提算法的可靠性及较高的检测精度。
[Abstract]:With the acceleration of urbanization in China and the sustained rapid growth of motor vehicle ownership in cities, traffic problems are becoming increasingly serious. The parking survey data of large and medium-sized cities in China show that the problem of roadside parking is especially serious, especially during the peak period, the dynamic traffic congestion is serious. In this context, the Intelligent Transportation system (ITS) is gradually formed. In intelligent transportation system, traffic information collection (such as vehicle detection) plays an important role, which is the basis of traffic flow prediction, control and rapid response to emergencies. Because the vehicle is a ferromagnetic material, the presence of the vehicle in space will cause disturbance to the geomagnetic field, so the disturbance can be detected by the anisotropic magnetoresistive (AMR) sensor on the ground, thus achieving the purpose of vehicle detection. The AMR sensor node is connected to the wireless network, and the (WSN), can be widely used in the intelligent transportation system. At present, the vehicle detection technology based on wireless magnetoresistive sensor network is not mature, and the basic problems of vehicle parking detection are still not well solved. In this paper, the above problems are studied. Firstly, the characteristics of the signal and its identification method are analyzed through the extraction and investigation of the disturbance signal of vehicle parking to the geomagnetic field; secondly, the characteristic data of the signal are extracted. Based on the existing vehicle detection algorithms, a local extremum detection algorithm (REA),) is proposed based on the existing vehicle detection algorithms, and the cooperative decision detection algorithm (CDA).) for multi-node detection is proposed by using cooperative information processing strategy. The local extremum detection algorithm adopts a process-based method. Considering the changing process of parking signal, the fluctuation feature of the signal is extracted in real time by the state machine, and the corresponding judgment rules are designed according to the characteristic data. The cooperative decision detection algorithm focuses on the correlation of the adjacent parking space signals, and triggers the detection at the sensor node with a small amplitude. According to the characteristic data obtained by the REA algorithm, The router fuses the information of different nodes to make a comprehensive judgment. The two algorithms are applied to sensor node and router node respectively. The algorithm has been applied in the actual system for more than 6 months and the number of nodes is over 100. The reliability and high detection accuracy of the proposed algorithm are proved by the experimental verification and the feedback data of the actual system.
【学位授予单位】：南京邮电大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：U495;TP212.9;TN929.5

【参考文献】