Real-time estimation method for turning proportion of level crossing of urban roads

The invention discloses a real-time estimation method for the turning proportion of a level crossing of urban roads. The method comprises that sectional traffic flow data of an entrance lane of the crossing and traffic flow data of a downstream road corresponding to an exit lane of the crossing are obtained, the flow deviation mode between the exit lane and the downstream road thereof is analyzed, and a Kalman filtering model for turning proportion estimation is established. According to flow conservation of the entrance and exit lanes of the crossing and random roaming of flow deviation, an observation equation of the Kalman filtering model for turning proportion estimation is established; according to the continuity and smoothness of short-period fluctuation of the turning proportion, a state equation of the Kalman filtering model is established; and the turning proportion of the crossing is estimated in real time by utilizing the Kalman filtering algorithm. The substantial random fluctuation mode of flow deviation between the exit lane of the crossing and the downstream road of the exit lane is fully dug, and accurate real-time turning proportion information of the traffic flow of the crossing can be provided for design and optimization of crossing signals of the urban roads.