面向交通信息采集的无线磁阻传感器网络协同感知研究

In recent years, with the rapid growth of car ownership, the traffic congestion problem is becoming increasingly serious in the center of the city. So there is an urgent demand for construction of the real-time traffic guidance and roadside parking management system. The project is based on wireless magnetic sensor network (WMSN), by use of geomagnetic interference of vehicle detection principle, to build large-scale, low-cost, low-power dynamic and static traffic detection system. Detection accuracy of the existing systems is not high. Currently, research on related laws and fusion methods of magnetic signal are inadequate. The project makes use of theoretical and statistical analysis methods to in-depth study the related laws of the magnetic signal. Expand collaborative awareness of time and space coverage, in-depth analysis of signal correlation between sensor nodes. Explore new features for accurate detection of vehicle signal. And we will propose energy efficient collaborative information fusion method. The fuzzy results will through the multi-level collaborative information fusion to reduce false detection rate. Using Bayesian statistical and CART decision trees methods to implement dynamic and static traffic detection algorithm respectively. Finally, we'll construct a traffic detection system to verify the validity of the project's research results. Research in this project will significantly improve the accuracy of traffic detection, and provide a new technical approaches for the city traffic information collection.

随着机动车保有量的高速增长，交通拥堵问题日益严重，各大中城市对实时交通诱导及路边停车管理有迫切的建设需求。基于无线磁阻传感器网络，利用地磁扰动车辆检测原理，可构建大规模、低成本、低功耗的城市道路动态及静态交通全覆盖的检测系统。现有系统检测精度不高，节点间磁信号的关联规律及融合方法研究不足。本项目利用磁场建模和统计分析相结合的方法对磁信号关联规律进行深入研究，扩大协同感知的时间和空间覆盖范围，分析深层次信号间关联规律，寻找对车辆信号精确检测的新特征；提出能量高效的协同信息融合方法，将模糊检测结果通过多级协同信息融合减少误判率；利用贝叶斯统计及CART决策树方法分别实现动态及静态交通检测算法。并建立道路动态及静态交通检测系统以验证理论研究的有效性。本项目的研究将显著提高磁阻传感器网络交通检测精度，为城市的交通信息采集提供新的技术途径。

随着机动车保有量的高速增长，交通拥堵问题日益严重，各大中城市对实时交通诱导及路边停车管理有迫切的建设需求。基于无线磁阻传感器网络，可构建大规模、低成本、低功耗的城市道路动态及静态交通全覆盖的检测系统。本项目研究磁场建模，对磁信号关联规律进行深入研究，分别实现了动态及静态交通检测算法。（1）根据车辆磁偶极子模型推导了邻近车辆信号相关规律；（2）提出了一种新的基于距离判别方法的停车检测算法；（3）提出了一种基于动态时间规整方法的动态车辆检测方法；（4）提出了一种动态路由系统，可自适应的进行网络拓扑变化。并建立道路动态及静态交通检测系统验证了理论研究的有效性。本项目的研究提高了磁阻传感器网络交通检测精度，为城市的交通信息采集提供新的技术途径。