Optimal control strategy for vehicle starting coordination based on driver intention recognition

Abstract

To enhance the starting performance of heavy-duty vehicles under different starting conditions, a vehicle starting coordinated optimal control method based on driver intention recognition is proposed. This method uses the Gaussian Mixture Model-Hidden Markov Model (GMM-HMM) for starting intention recognition, dividing the starting intentions into three categories: gentle start, normal start, and emergency start. The GMM-HMM starting intention recognition model is validated using real vehicle data. Based on the recognition results of driver intentions, a performance index function is defined as a weighted sum of smoke limit restriction time, 0–20 km/h acceleration time, and starting jerk. By assigning different weight coefficients, the allocation of requirements for starting power and comfort is achieved. Based on the principle of minimizing values, the coordinated control parameters (upshift speed and starting fuel quantity) are optimized, resulting in the optimal combination of coordinated control parameters under different starting intentions. This enables the optimal control of vehicle starting coordination based on the driver’s different starting intentions.