基于多传感器数据融合技术的短时交通流检测  

Intersection traffic volume detection based on multi-sensor fusion technique

作　　者：徐华中[1] 吴苏[1] 刘念[1] 

机构地区：[1]武汉理工大学自动化学院,湖北武汉430070

出　　处：《传感器与微系统》2009年第2期104-106,109,共4页Transducer and Microsystem Technologies

摘　　要：在阐明智能交通控制系统(ITS)数据融合的意义和层次性的基础上,分析了数据层多源数据融合和神经网络的特点,根据遗传算法与BP算法各自的优缺点,设计了将遗传算法与改进的BP算法相结合的方法应用于融合计算,并提出了该方法的实现步骤。利用这种方法,对交叉路口的多源传感器数据进行融合计算后,仿真结果证明:该方法大大减少训练时间,能够有效地进行数据质量控制,提高数据的精度。Through a characteristic analysis of multi-sensors intelligent transport system（ITS） data fusion on data layer and neural network,a multi-sensors ITS data fusion approach using the genetic algorithm and the improved BP networks algorithm on the base of their advantage and disadvantage is proposed, and the implementation processes is designed. The data fusion approach applied to eight-flow data from crossway demonstrates that the proposed fusion approach can not only save the training time greatly, but also process the data quality control effectively and improve the level of the data accuracy.

关 键 词：数据融合 交通量检测 遗传算法 神经网络 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]