A Novel Prediction Method of Optimal Driving Speed for Intelligent Vehicles in Urban Traffic Scenarios

Due to sensitively reflect traffic conditions in road network, driving speed is a vital index to evaluate the capability of urban traffic network and intelligent transportation system. The research on optimal driving speed is beneficial for improving the stability and safety of intelligent vehicles in urban traffic scenarios. This paper puts forward a novel prediction method of optimal driving speed for urban intelligent vehicles based on road design principles and traffic flow theories. Firstly, driving factors of urban traffic scenarios are selected, quantized and stored in the optimal-driving geographic information database established for intelligent vehicles. Secondly, multivariate linear equations are investigated using regression analysis methods to reveal the relationship between driving speed of intelligent vehicles and related factors, including urban road parameters, real-time traffic conditions and vehicle information. Thirdly, urban traffic variable-length model is built to explore dynamic characteristics of traffic evolution, further macroscopic constraint equations of driving speed for intelligent vehicles are deduced based on traffic fundamental diagram. Finally, regarding instant travel time, total travel time and total travel distance as evaluation metrics, the optimal solution of the multivariate linear equation and macroscopic constraint equations is calculated ultimately, which is the optimal driving speed for intelligent vehicles in the urban traffic network. Simulation results have proved that the proposed prediction method can provide safe, feasible and efficient driving speed advice for intelligent vehicles in urban traffic scenarios.