短时交通流预测方法比较  

The Discussion of Methods of Short-term Traffic Flow Forecast

作　　者：唐丽娜[1] 张卫华[1] 

机构地区：[1]合肥工业大学交通研究所,安徽合肥230009

出　　处：《交通科技与经济》2008年第2期100-102,共3页Technology & Economy in Areas of Communications

摘　　要：短时交通流预测是实现交通流诱导的关键技术之一。短时交通流因为其不确定性等特点而使其预测很复杂。通过实地调查获取的交通流量数据,分别采用移动平均法、指数平滑法、AR模型法3种交通流预测方法进行短时交通流量预测,并通过不同的评价指标对上述3种方法的预测效果进行评价,得出AR模型方法的预测效果优于其他2种方法。Short-term traffic flow is one of the core technology to realize traffic flow guidance. It is complicated to forecast the traffic flow because of its uncertainty. Through the data obtained by fieldwork, the paper forecasts the short-term traffic by three methods： moving-average method, index-smoothing method, AR model method. Finally, we evaluate the three effects of the three methods through different evaluate targets and get the result that AR model is the better method.

关 键 词：智能交通系统 短时交通流预测 移动平均法 指数平滑法 AR模型法 

分 类 号：U491.112[交通运输工程—交通运输规划与管理] TP317.3[交通运输工程—道路与铁道工程]