Multi-Objective Vehicle Speed Planning with Incorporation of Driving Style Classification

Effective identification driving styles of drivers for vehicle speed planning can improve vehicle fuel economy, driving adaptability and comfortability. First, based on historical driving data, a driving style classification and identification framework using improved k-means algorithm and support vector machine is proposed in this paper, which classes and identifies drivers into three groups: conservative, regular, and aggressive. Then, on the premise of ensuring vehicle traffic efficiency, a multi-objective speed planning optimal control problem is constructed for three types of drivers in the intelligent transportation system to achieve preferable fuel economy, driving adaptability, and driving comfort. The problem is then discretized into the nonlinear programming problem for solving by orthogonal collocation direct transcription algorithm. Finally, the proposed multi-objective speed planning framework is applied to simulate the speed of drivers with different styles in a speed planning scenario, and the speed planning results are compared with those without driving styles. The simulation outputs indicate that the constructed framework can effectively improve the driving adaptability, comfort, and reduce the fuel consumption of different driving styles while ensuring the traffic efficiency.