基于反馈机制的城市道路短时交通流量预测研究

Short-term Traffic Flow Prediction for Urban Streets Based on Feedback Mechanism

交通精细化管理对掌握交通流规律提出了更高要求,交通流量预测是提高交通控制与管理水平的重要保证。针对城市道路短时交通流量预测忽视运用实时检测数据对预测模型进行反馈的问题,文中提出由反馈模块和预测模块2部分组成的预测方法,并运用上海市数据对方法的有效性进行验证。结果表明,本方法的平均、绝对、相对误差均小于基于反馈机制的单一模型及没有反馈机制的单一模型或组合模型,有效地提高了城市道路短时交通流量的预测精度。

Fine management of transportation has put forward higher requirements for grasping the law of traffic flow,and traffic flow prediction is an important guarantee for improving traffic control and management level.A traffic flow prediction method is proposed,which consists of two modules,namely feedback module and prediction module.Traffic flow data in Shanghai is used to verify the effectiveness of the method.The results show that the average,absolute and relative errors of proposed method for all intervals are less than those of the single model based on feedback mechanism and the single or hybrid model without feedback mechanism.The proposed method can effectively improve the accuracy of short-term traffic flow prediction.