Prediction models of freeway traffic incident delay

Abstract

Incident impact on the road traffic directly is to cause congestion and delay. Firstly, certainty queuing model combined with variables of four stage duration prediction was employed to develop the incident certainty delay model. And the stochastic delay real-time prediction model was created combining the random distribution characteristics. With the time information when vehicles entered and exited the toll station, we could get the actual value of incident delay. Comparing with the estimated delay and the actual delay, the test results indicated that the estimated models were reliable and applicable.