Optimization of New Energy Vehicle Transmission System based on Feedforward Neural Network

Optimization of new energy vehicle (NEV) transmission systems is crucial in order to promote energy efficiency, driving performance, and sustainability. Conventional rule-based control systems tend to be suboptimal because of vehicle nonlinear dynamics and varying driving conditions. Although techniques such as fuzzy logic and model predictive control can provide improvements, they are hampered by intensive computational requirements and lack of adaptability. This paper presents an FNN-based optimization model that is capable of addressing these issues by learning intricate nonlinear relationships and facilitating real-time optimization. The FNN model is trained on varied driving condition data to optimize gear shift timing and torque distribution, leading to improved performance and energy efficiency. Comparative analysis indicates that the FNN model obtains better results, with an accuracy of 97% and an F1-score of 97.2%, outperforming conventional rule-based approaches. Moreover, the suggested FNN model has lower computation cost and greater predictive accuracy compared to conventional machine learning techniques, thus being suitable for real-time applications. The results indicate the efficacy of neural network-based optimization in NEVs, providing better energy efficiency, longer battery lifespan, and increased driving comfort.