Velocity Prediction of Intelligent and Connected Vehicles for a Traffic Light Distance on the Urban Road

Accurate vehicle velocity prediction has important theoretical value and widespread applications in many areas, such as optimal control of vehicle propulsion system, eco-driving, and advanced driver assistance systems. However, for dynamic changes of traffic condition caused by traffic lights, intersections, and other factors, it is hard to predict the vehicle velocity accurately on the urban road. In this paper, we present a novel vehicle velocity prediction algorithm for intelligent and connected vehicles based on the historical driving data of the preceding vehicle and traffic light information. First, the basic driving rules on the urban road are studied in two different driving scenarios. Then, a vehicle trajectory generation algorithm (VTGA) is proposed to generate the vehicles' trajectories according to the basic driving rules. To identify vehicles' quantity and the global positioning system information of each vehicle in the unknown area, an identification algorithm (IA) is designed based on the VTGA and genetic algorithm. Finally, a vehicle velocity prediction algorithm is applied to predict the velocity of the target vehicle based on the VTGA and the results of IA. To verify the method proposed in this paper, the next generation simulation database is utilized. The results demonstrate that the accuracy of the vehicle velocity prediction has a significant improvement in the urban network, and the root-mean-square error reduces from 0.50 ~ 4.78 m/s (5 ~ 20 s) to 0.7594 ~ 0.9166 m/s (9.3 ~ 43.8 s), when compared with methods of other studies.