城市快速路交通事件检测的自适应算法研究  

Incident Detection in Urban Freeway Traffic Based on Adaptive Algorithm

作　　者：田秋芳 陈阳舟[1] 张利国[1] 

机构地区：[1]北京工业大学电子信息与控制工程学院,北京100124

出　　处：《交通运输工程与信息学报》2010年第4期99-103,125,共6页Journal of Transportation Engineering and Information

基　　金：国家自然科学基金(60904069)

摘　　要：针对城市快速路交通事件检测问题,提出了一种基于自适应遗传算法与神经网络相结合的自动检测算法。通过改进的自适应遗传算法优化神经网络结构和权值参数,保证了神经网络能以较小规模和最优的权值参数来描述事件发生与交通参数间的映射关系,从而提高检测效果。利用PARAMICS交通软件模拟了北京市京通快速路从大望桥到四惠桥路段间的一组交通数据,仿真结果表明,该算法同现有的典型算法相比较,具有较高的检测率和较快的检测速度。An automatic incident detection algorithm based on an adaptive genetic algorithm combined with neural networks was proposed for urban freeway traffic.The neural structure and parameters were optimized by the adaptive genetic algorithm,which could describe the relation between traffic incident and the traffic parameters and improve the detection results.A set relevant traffic data was obtained with PARAMICS traffic software by simulating the road section between DAWANG bridge and SIHUI bridge of Jingtong freeway in Beijing city.The simulation results indicated that the given algorithm has better detection rate and detection speed for urban freeway traffic,compared with some classical detection algorithms.

关 键 词：城市快速路 交通事件检测 自适应遗传算法 算法研究 Adaptive Algorithm Based 神经网络结构 遗传算法优化 交通参数 映射关系 软件模拟 权值 交通数据 检测效果 检测问题 检测算法 检测速度 仿真结果 PARAMICS 小规模 

分 类 号：U[交通运输工程]