Mobile Phone Data Feature Denoising for Expressway Traffic State Estimation

Due to their wide coverage, low acquisition cost and large data quantity, the mobile phone signaling are suitable for fine-grained large-scale estimation of traffic conditions. However, relatively high level noise makes it difficult achieve sufficient accuracy. According characteristics noise, this paper proposed an improved density peak clustering algorithm (DPCA) filter noise. In addition, on basis long short-term memory model (LSTM), a state based feature was established with use denoising realize expressway precision, fine granules, coverage. The Shanghai–Nanjing Expressway used as case study area method verification, results which showed that in can effectively reduce impact extreme data, significantly improve accuracy state, reflect actual situation fairly satisfactory manner.