An adaptive traffic parameter estimation method based on hybrid particle filter for rainy weather

In order to improve the estimation accuracy of highway traffic parameters, on the basis of the macro traffic flow model and the state-space model, this paper proposes a traffic parameter estimation method based on hybrid particle filter according to the Bayesian theory. Considering the sensitivity of the estimation results to the changes of the model parameters, and to avoid the impacts from the preset fixed model parameters on the estimation accuracy, this paper establishes the relationship between the free flow velocity and the traffic condition (saturation V/C) and proposes an adaptive adjustment strategy for model parameters under the effect of traffic conditions. The test results show that: the traffic parameter estimation based on hybrid particle filter is much more accurate than the Kalman filtering estimation and can quickly identify the obvious fluctuations in traffic volume, exhibiting its strong stability in the rainy weather. In addition, the adaptive adjustment strategy for model parameters under the influence of traffic conditions can significantly improve the accuracy of the traffic parameter estimation, and can still provide fairly good estimation results even in the event of an accident.