Travel Time Probability Prediction Based on Constrained LSTM Quantile Regression

Travel time reliability assessment has been widely used in recent years to evaluate the performance of transportation networks and measure operation level systems. Weather, as one important factors influencing travel reliability, affects relationship between supply requirement urban road networks. Considering traffic characteristics under different conditions, a study on influence weather conditions is proposed predict probability travelers completing their trips within expected conditions. Based network data cab trajectory Harbin city, this paper correlates floating vehicle location with information through hidden Markov model reduce errors calculation results path time. To analyze entire distribution extreme its impact various situations, it captures tail features based value theory. Then, increase predictability each quantile, combines deep-learning LSTM quantile regression create probabilistic prediction utilizing combined layers. The compared linear neural models, evaluated terms point results, respectively, ensure accuracy predictions from model. As result, greatly improved, degree violating constraints reduced.