Methods of real-time screening and reconstruction for dynamic traffic abnormal data

Cheng XU,Zhaowei QU,Pengfei TAO,Sheng JIN

DOI: https://doi.org/10.11990/jheu.201503045

2016-01-01

Abstract:Because of the large quantity of absence and abnormity of original traffic data in the real detector, by an?alyzing the influence of inner relationship of data acquisition interval and three traffic flow parameters and other fac?tors, this paper designed a four?step data sieving method, including preliminary screening, threshold filtering, the?oretical screening of traffic flow and quality control sieving. Then four data reconstruction methods used in different conditions were also proposed based on time series, historical data, spatial location and spatial?temporal correla?tion, lastly on the basis of this, the data preprocessing process was set up. Using the field data on the city express?way in Beijing, it indicates the algorithm can effectively remove abnormal data and the abnormal data reconstruction accuracy is less than 10%. These methods have been proven to have good real?time performance and stability, and can be used in engineering projects.