Traffic Measurement and Vehicle Classification with Single Magnetic Sensor
Traffic Measurement and Vehicle Classification with Single Magnetic Sensor
复制标题
DOI：
10.1177/0361198105191700119
复制
发表时间：
2004-09
影响因子：
1.7
通讯作者：
S. Cheung;S. Coleri;B. Dundar;Sumitra Ganesh;Chin-Woo Tan;P. Varaiya
中科院分区：
文献类型：
--
作者：
S. Cheung;S. Coleri;B. Dundar;Sumitra Ganesh;Chin-Woo Tan;P. Varaiya
无线磁传感器网络为高速公路和十字路口的交通测量提供了一种有吸引力的，低成本的替代感应回路。除了提供车辆计数、占用率和速度外，这些传感器还可以提供无法从标准环路数据中获得的信息（例如非基于轴的车辆分类）。因为这样的网络可以快速部署，所以它们可以用于（和重用）临时流量测量。本文在两个野外实验的基础上，对磁传感器的探测能力进行了研究。第一个实验在加州伯克利的赫斯特大道上收集了2小时的测量轨迹。车辆检测率高于99%（摩托车以外的车辆检测率为100%），车辆平均长度和速度的估计值似乎高于90%。测量还产生车间距或车头时距，揭示有趣的现象，如队列形成下游的交通信号。第二个实验的结果是初步的。对同一地点37辆过往车辆的传感器数据进行处理并分为六种类型。当不使用长度作为特征时，60%的车辆被正确分类。该分类算法可以由传感器节点本身在真实的时间内实现，与其他基于高扫描速率感应回路信号的方法相比，这些方法需要大量的离线计算。据信，如果使用长度作为特征，80%至90%的车辆将被正确分类。
Wireless magnetic sensor networks offer an attractive, low-cost alternative to inductive loops for traffic measurement in freeways and at intersections. In addition to providing vehicle count, occupancy, and speed, these sensors yield information (such as non-axle-based vehicle classification) that cannot be obtained from standard loop data. Because such networks can be deployed quickly, they can be used (and reused) for temporary traffic measurement. This paper reports the detection capabilities of magnetic sensors on the basis of two field experiments. The first experiment collected a 2-h trace of measurements on Hearst Avenue in Berkeley, California. The vehicle detection rate was better than 99% (100% for vehicles other than motorcycles), and estimates of average vehicle length and speed appear to have been better than 90%. The measurements also yield intervehicle spacing or headways, revealing interesting phenomena such as platoon formation downstream of a traffic signal. Results of the second experiment are preliminary. Sensor data from 37 passing vehicles at the same site are processed and classified into six types. Sixty percent of the vehicles are classified correctly when length is not used as a feature. The classification algorithm can be implemented in real time by the sensor node itself, in contrast to other methods based on high-scan-rate inductive loop signals, which require extensive off-line computation. It is believed that if length were used as a feature, 80% to 90% of vehicles would be correctly classified.