Vehicle spatial and time trajectory filling based on dynamic road network

Complete vehicle trajectory data is essential for urban traffic flow modeling studies. This study proposes a framework for filling vehicle trajectories in spatial and time for automatic vehicle identification (AVI) data. Based on the particle filter, the dynamic correction factor is innovatively used to improve algorithm accuracy. After four resamplings, such as traffic situation index and traffic event factor, spatial trajectory filling is completed. The Copula function fills the time trajectory by analyzing the correlation between upstream and downstream paths. Finally, the experiment was conducted in Xiaoshan District, Hangzhou, China. The results show that for spatial trajectory filling, the average accuracy exceeds 97% with 75% camera coverage. In time trajectory filling, the time trajectory filling error is reduced by 35% compared to the Hellinga algorithm.