Real-time travel time estimation in partial network coverage: A case study in Kuwait City

Travel time is an important metric for both road users and transportation professionals. Collecting travel time in large road networks is costly and faces the challenge of partial network coverage resulting from an inadequate number of sensors (moving or fixed sensors). This study proposes a solution to the problem of limited network sensor coverage caused by insufficient coverage of probe vehicles or limited fixed sensors. The study uses travel time correlation and the concept of neighbor links to develop dynamic models that are capable of estimating real-time travel times on links not covered by sensors in Kuwait City. Neighbour links are a set of nearby links that have similar traffic characteristics, located in the same area type, and their travel times are correlated. The performance of the developed models is validated using field-measured and simulated travel time data on 70 links in Kuwait City. VISSIM microsimulation model is developed to investigate the impact of probes (moving sensors) penetration rate on travel time estimation accuracy and links Coverage. The results show that the dynamic model is feasible, applicable and provides a satisfactory accuracy even at low probe penetration rate of 1%. However, based on links coverage, a probe penetration rate of 3% is recommended. The average links coverage reached 86.1% at a penetration rate of 3%. This research can find applications in Advances Traveler Information Systems, traffic planning, signal design, and demand management.