基于RBF算法的公交车到站时间预测

Bus arrival time prediction based on RBF algorithm

传统的公交车时间预测由于多方面因素影响导致很难建立数学模型,预测结果精度不理想很难让市民满意,这将很难提高公共交通的服务水平和服务质量。由于RBF神经网络能够迅速地逼近实际值,为此文中提出一种基于RBF神经网络的公交车到站时间预测,以公交车到站时间为输入,以两站相差的时间为输出建立模型。通过仿真和分析,并且和BP神经网络进行预测结果对比,实验结果表明,RBF神经网络在公交车到站时间预测结果与实际值更为接近,并且优于BP神经网络。

The traditional bus time prediction is hardly possible to establish a mathematical model because of the various factors,and the prediction accuracy is not ideal and it is very difficult to achieve public satisfaction,which will be hard to improve the service level and service quality of the public transportation.As the RBF neural network can approach the actual value quickly,a bus arrival time prediction based on RBF neural network is proposed,which takes the bus arrival time as the input and takes the time difference between two stations as the output to modelling.The simulation and analysis results show that,in comparison with the BP neural network,the predicted results of RBF neural network in the bus arrival time are closer to the actual values,and better than that of the BP neural network.