基于Fisher判别的城市快速路自动事件检测算法研究

交通事件检测是尽早发现公路运营中的事故隐患,提高运输安全性的重要技术.文章目的在于构建一个高精度并且易于实施的快速路自动事件检测算法.在简述国外自动事件检测算法、分析快速路特点和事件检测系统要求的基础上,研究了快速路检测器布置的原则和优化密度,选择了对事件发生敏感度最大的4个交通参量,建立了基于Fisher判别模型的自动事件检测算法,采用历史数据对模型进行标定,并给出了试验评价结果.试验结果表明,该方法检测率达到85.7%,并且检测速度快.

Automatic Incident Detection Algorithm Based on Fisher Discriminant for Urban Expressway

Traffic incident detection is an important and necessary technique to ensure that all incidents are found as early as possible and thus improving highway transportation safety.The purpose of the paper is to put forward a high precision Automatic Incident Detection(AID) model for urban expressway easily for application.After an extensive literature review of AID techniques abroad,the paper analyzes both the characteristics of urban freeway and the critical requirements of AID system,and addresses traffic detectors layout as well as their optimization density for AID data collection.Further more,the paper presents an AID algorithm based on Fisher discriminant with selecting four traffic flow measurements most sensitive to incident occurrence.In the end,the AID model is tested with historical data collected from urban expressway,the evaluation results show that the AID model based on Fisher discriminant has perfect performance of fast detection with high detection rate up to 85.7%.