交通异常贝叶斯动态检测方法

The method of incident detection with dynamic bayes prediction

针对卡尔曼滤波方法在稳定性、滤波器发散及计算方面的缺陷，以占有率为观测变量，以交通流的密度和速度为系统状态，建立了交通异常贝叶斯动态检测模型及算法，能够实施无初始输入的异常检测。

The bayes dynamic detection model and algorithm are built in accordance with the defects of kalman filtering method in stability,filter divergence and calculation by taking occupancy as detecting variation and traffic flow density and speed as state of system,incident can be detected without initial input.