论文部分内容阅读
为研究适合自适应信号控制系统的流量预测模型,利用ARIMA模型进行数据预处理的基础上,考虑高阶神经网络收敛速度慢及易陷入局部最小点的特点,通过在线调整学习率及引进动量法对其进行改进,得出基于ARIMA与改进的高阶神经网络的组合预测模型,试验表明预测的交通流量满足自适应信号控制系统实时在线多频优化的时间及精度要求.
In order to study the traffic prediction model suitable for adaptive signal control system and to use ARIMA model for data preprocessing, considering the characteristics of high-order neural network’s slow convergence speed and falling into local minimum, by adjusting the learning rate online and introducing momentum method The improved prediction model based on ARIMA and improved high-order neural network is obtained. The experimental results show that the predicted traffic can meet the time and accuracy requirements of real-time on-line multi-frequency optimization of adaptive signal control system.