基于粒子群优化的支持向量回归车道饱和度预测

Traffic Lane Saturation Prediction with the Support Vector Regression Based on Particle Swarm Optimization

针对复杂城市交通系统,建立基于粒子群优化的支持向量回归车道饱和度预测模型。首先,引入悉尼自适应交通控制系统中的交通数据,对路网的交通运行状况以及路口之间的关联关系进行分析,利用均值滤波方法对数据进行预处理,确保数据的可靠性。然后,基于支持向量回归方法建立车道饱和度预测模型,并引入粒子群优化算法对模型的参数值进行寻优,从而得到参数最优的车道饱和度预测回归模型。最后,对基于实际数据建立的模型进行验证。验证结果表明,预测数据与实测数据的拟合程度较高,所建立的预报模型能够有效预测将来车道饱和度信息,以及可能发生的道路拥堵。

Aiming at the complex urban traffic system,a prediction model of urban traffic lane saturation is built with support vector regression based on particle swarm optimization (PSO).Firstly,the data collected from Sydney adaptive traffic control system (SCATS) is used to analyse the relationships of traffic operation status and each intersection in the road network.The mean filtering method is used to process the traffic data to ensure the reliability of data.Then,the traffic lane saturation prediction model is established based on support vector regression method;and particle swarm optimization algorithm is used to optimize the parameters of the model,thus the optimal regression prediction model of traffic lane saturation is obtained.Finally,the model established based on actual data is verified.The results show that the predicted data fitted well with the real data,the established prediction model can be used to effectively predict the future traffic lane saturation and possible road congestion.