城市交通信号的ANN自校正预测控制  

Artificial neural networks self-tuning predictive control for traffic signals

作　　者：刘智勇[1] 李水友[1] 

机构地区：[1]五邑大学304信箱,广东江门529020

出　　处：《控制理论与应用》2003年第6期933-937,共5页Control Theory & Applications

基　　金：教育部高等学校骨干教师资助计划项目(教技司[2000]65号);广东省自然科学基金项目(010486).

摘　　要：提出一种基于人工神经网络的城市交通信号的自校正预测控制方法.充分考虑相邻交叉路口之间交通流的强耦合性,在此基础上建立关于队长的交通模型;其中,受控路口下一周期到达的车辆数用人工神经网络(ANN)来预测;通过该ANN还可获得确定最佳周期长度所需要的交通参量,因此还可预测下一周期的长度;上述预测值均用实测信息进行反馈校正,在此基础上即可给出带约束的预测控制算法,从而确定下一周期的控制策略.仿真实例表明该方法具有较好的控制效果.A self-tuning predictive control method for traffic signals in urban area based on artificial neural networks (ANN) was proposed. The strong coupling between adjacent signaled intersections was fully considered, based on which a traffic model for queues was established. The number of approaching vehicles in the next cycle at this intersection was predicted by using ANN and the traffic parameters were obtained to decide the optimal cycle length. The feedback tuning method was used to process all the above prediction values by using real measured data. The constrained predictive control algorithm was given to determine the control tactics. Simulation results showed that the proposed method is effective.

关 键 词：城市交通控制 智能交通系统 城市交通信号 ANN 自校正预测控制 人工智能 人工神经网络 

分 类 号：TP273.5[自动化与计算机技术—检测技术与自动化装置]