Traffic Flow Long-term Forecasting Based On Type-2 Fuzzy Set

With the rapid development of data detection technology and data storage technology, how to handle the huge amounts of obtained traffic flow data effectively has become the main focus in intelligent transportation systems. In the field of traffic flow data processing, long-term traffic flow data forecasting is of great significance. This paper adopts type-2 fuzzy sets theory to deal with measured traffic flow data and accomplish long-term forecasting due to its capability of dealing with random and uncertain data. The main work of this paper is demonstrated as follows.Firstly, this paper proposed a long-term traffic flow forecasting frame and its implement method based on interval type-2 fuzzy set. The frame was divided into three parts including traffic flow data preprocessing module, type-2 fuzzification module and long-term traffic flow data module.Secondly, this paper applied interval type-2 set in traffic field based on type-2 fuzzy sets theory and central limit theory. Central limit theory converts point data into confidence intervals, which provides the basis of application in other fields. Besides, the long-term forecasting scheme was able to deal with big traffic flow data because randomness and uncertainty of the data can be retained effectively. The scheme was proved to be equipped with good fault-tolerant ability even without data cleansing.Thirdly, the long-term traffic flow forecasting frame was proved feasible and valid through a series of simulations. Different from conventional long-term forecasting method, the result of this paper is a region surrounded by two forecasting curves, which is able to express the range of traffic flow fluctuation effectively. Besides, this paper discussed the influence of scale factor and filtering methods.