Prediction of the shortest travel time based on intersection delay

Abstract

Based on positioning data of more than 4000 floating cars and GIS map in Xi'an, the prediction of the shortest travel time based on intersection delay is proposed in urban complex environment. GIS map and floating car data (FCD) are extracted and cleaned by preprocessing. An improved map matching method is put forward, which is more efficient and accurate. With the consideration of the instantaneous road speed and road intersection delay, the shortest travel time will be predicted. The low searching efficiency, caused by the urban road network nodes, is overcome by dynamically generating the road network matrix with searching points. The algorithm has been realized and tested. The results show that the map matching efficiency is much higher and correct rate is 92.1%. The searching efficiency can be increased more than 50% by dynamically generating the road network matrix. Travel time prediction is matching more 90% with the actual time, and it can save about 20% of the original travel time under normal urban traffic conditions. The study can adapt to the complex urban road traffic conditions, greatly reducing people's travel time.