基于径向基函数神经网络的交通信号控制优化
Traffic Signal Control Optimization Based on Radiationl Basis Function Neural Network

文章采用迭代训练的学习算法，结合径向基函数构建神经网络，通过构建的网络能够得到交通信号控制策略，为交叉路口的交通问题提供更加高效的交通信号控制方案。文章的神经网络与传统的交通控制策略相比，降低了车辆的平均延误时间。针对移动交通流检测信息的特点，提出了一种基于移动交通流检测信息的路况概率神经网络判别方法，通过分析路况的相关因素，同时考虑信号控制交叉口红灯对车辆行程时间延误的影响，通过结合径向基函数算法改进的概率神经网络，根据实地调查的数据，得出预测更加准确的神经网络。预测的结果通过对比，降低了交叉路口车辆的平均延误时间，减少了交通安全隐患。
In this paper, the learning algorithm of iterative training combined with Radiationl Basis Function (RBF) is used to construct a neural network, through which the traffic signal control strategy can be obtained to provide a more efficient traffic signal control scheme for the traffic problems at intersections. Compared with the traditional traffic control strategy, the neural network in this paper reduces the average delay time of vehicles. According to the characteristics of mobile traffic flow detection information, a road condition probabilistic neural network discriminant method based on mobile traffic flow detection information is proposed, by analyzing the relevant factors of road conditions, and also considering the influence of the signal control intersection red light on the travel time delay of the vehicle, by combining the RBF algorithm to improve the probability neural network, according to the field survey data, the prediction of the neural network is derived more accurately. By comparing the predicted results, the average delay time of vehicles at intersections is reduced, and the hidden danger of traffic safety is reduced.