Study On The Methods Of Short-term Traffic Flow Forecasting Based On RVM And ARIMA

The urban traffic congestion problems have become more serious with continuously push forward the process of urban modernization and the continuous growth of vehicle ownership, it greatly increases the cost and travel time of residents,so it has become a social problem. Intelligent transportation systems are considered as one of effective methods to relieve the urban traffic problems such as traffic congestion, air pollution caused by motor vehicle exhausts and traffic accidents. Short-term traffic flow forecasting is not only a core element of intelligent transportation systems but also an important base of traffic information service, traffic control. It also can be used as a key basis for traffic decisions, while can supple travelers with efficient information and help them to choose an optimal path. Therefore, the research of short-term traffic flow forecasting is the basic work to accurately grasp the trend of change for traffic parameters.The purpose of this paper is to further improve the accuracy of short-term traffic flow forecasting. Firstly, this paper analyzed the synthesis of short-term traffic flow in different time scales on the basis of clearly the defined short-term traffic flow, and this paper also analyzed short-term mutagenicity, long-term trend and stochastic volatility characteristics with a combination of different time scales space-time diagram of short-term traffic flow. Secondly, a noise reduction method of short-term traffic flow which based on relevance vector machine(RVM) was established, and the step of noise reduction method of short-term traffic flow was designed. Then, this paper selected evaluation to test the smoothing method, and the effectiveness of the noise reduction method was confirmed by test function. Moreover, this paper analyzed the smooth and steady of short-term traffic flow of time series, and a method of short-term traffic flow forecasting which based on RVM and ARIMA was established. Besides, the step of method of short-term traffic flow forecasting was well designed, and mean absolute percentage error(MAPE) was selected as evaluating indicator of the method. Finally, the effectiveness of the method of short-term traffic flow forecasting was confirmed by the traffic flow of recording data in an urban road.Results indicated that the MAPE of method of short-term traffic flow forecasting was smaller than the MAPE of exponential smoothing model, BT neural network model and ARIMA model to forecast short-term traffic flow directly, and the proposed prediction method can effectively improve the accuracy of short-term traffic flow forecasting.