流量预测精度对控制效果的影响

Study on effect of flow forecasting error on traffic control performance

为说明流量预测算法对交通控制的实际影响,进而提出交通控制对预测算法的精度要求,根据青岛市江西路实地流量数据,分析线性预测算法的预测精度,研究预测精度对交叉口流量比、周期的影响,进一步获得预测精度对交叉口平均延误和通行能力的影响。研究过程发现数据前期的平滑处理可提高流量预测精度,从而减少预测数据对配时参数以及控制效果的干扰。通过数据分析及研究,结果显示线性流量预测算法预测误差大约为10%,这对周期造成2%左右的误差,使绿灯时间误差在2s以内,而对延误及通行能力的误差干扰都在5%以内。说明线性预测算法在实际信号配时方案中具有可行性,从而简化系统的复杂性,提高运行效率。

With many flow forecasting algorithm occur, there are few research on flow forecasting error influence on traffic control parameters and performance. So according to data of Jiangxi Road in Qingdao city, this paper analyses the feasibility and application of conventional linear forecasting model. This paper gets predicted error of the precision and estimates model’s interference on flow rate and cycle length, then evaluates the change on traffic control performance including average delay and capacity. In the analysis beforehand taking data smoothing can reduce forecasting error and lessen its effect on control parameters and performance. Finally the result shows that while forecasting error of the linear forecasting model is about 10% that causes cycle length 2% of forecasting error and the error green is in 2 second, average delay and capacity are less than 5%. Therefore the conventional linear model is feasible in practical signal timing and can simplify the complexity of system to improve performance and efficiency.