A Brief Review on Missing Traffic Data Imputation Methods for Intelligent Transportation Systems

Traffic data plays a vital action in transportation study and implantation, including forecasting travel times, designing travelling routes, and alleviating traffic congestion. When collecting traffic data from sensors, missing data issues are inevitable. How to perform accurate imputation based on incomplete traffic data has become a key research topic in Intelligent Transportation Systems (ITS). This technology is very challenging because of the complex nonlinearity and dynamic spatiotemporal correlation of traffic data. Recently, a large number of research efforts have significantly improved the performance of traffic recovery by modeling spatiotemporal traffic data through various methods. This paper provides a general summary of the literature on missing traffic data imputation from several aspects. Specifically, we discuss existing imputation methods, give classifications, and analyze current challenges and solutions.