Intelligent Driving Based on Multi-Objective Shifting Decision Method for Vehicle Transmission System

Effective gear shifting decisions are essential for ensuring vehicle safety, comfort, and fuel economy. This paper proposes an intelligent gear shift decision methodology for automatic transmission systems, utilizing advanced data mining techniques such as variational mode decomposition and data reconstruction. To enhance the accuracy of shift decisions, frequency synchronization processing and denoising are performed based on the variational mode decomposition method on the collected data, followed by the application of the compressed sensing orthogonal matching pursuit algorithm for handling missing values in the driving data. Subsequently, the optimal driving performance and optimal fuel efficiency shift curves are analyzed to determine the shift decisions. The moving least squares method is employed to model the relationship between vehicle speed, throttle position, and acceleration, capturing the shifting patterns of experienced drivers operating on straight and level roads. The experimental results show that when the throttle span is below 40%, drivers’ shift decisions tend to fall between the optimal fuel efficiency and performance curves, indicating a deliberate attempt to balance fuel efficiency with driving performance.