A Study for Highway Travel Time Prediction Using Markov Chain

抄録

Providing travel time information is not only basic part of ATIS(Advanced Traveler Information System) but also very important factor for user's route choice. However, although there are various method for collecting travel time data, time-lag have to be came up when data are collected. And it bring about low accuracy of information, especially on peak hour where information providing is more important. In this study, to estimate stochastic amount of time lag, using Markov Chain that show discrete probabilistic transition. Difference Ratio of TCS between actual travel time were used as transition probability of Chain Process. Through this process, gap between actual travel time and provided travel time became reduced. To verify this model, actual travel time data between Seoul IC and Suwon IC was applied into the model. Consequently, predicted data were verified to correspond with actual travel time by comparing with two-tailed t-test at 95% level of significant.