A Network-Wide Traffic Speed Estimation Model with Gaussian Process Inference

Paper in proceeding, 2023

Accurate urban road traffic speed analysis and prediction are important for the application of intelligent transportation systems. However, the limited and inefficient traffic state monitoring infrastructure installed on urban roads makes it difficult to monitor the traffic state of an entire network. Moreover, the complex characteristics of urban road networks may lead to difficulties for traditional statistical and traffic flow models in dealing with this type of complex relationship. Therefore, this study proposes a network-wide traffic speed estimation model with full spatial and temporal coverage and selects floating vehicle trajectory data in an actual road network for experiments. The results show that the proposed model can accurately estimate the full spatiotemporal traffic state of a traffic network with only partial data input. This method can be effectively applied to urban road state estimation and can provide a scientific basis for traffic management departments to formulate congestion mitigation strategies.