Development of an Electronic Differential System Based on Artificial Neural Networks for Electric Transport

The transmission of most modern vehicles consists of an internal combustion engine (ICE) transmitting force through a gearbox to a differential shaft which transmits torque to the wheels. In electric vehicles, electric motors can be attached directly to the wheels through a reduction gearbox. If the control system is appropriate, the vehicle can transmit different torques to the wheels on each side of the vehicle to ensure high roadability. This system is called electronic differential. Its implementation can be either very simple, based on the Ackerman-Jeantand model, which cannot fully replace a simple mechanical differential, or based on robust control mechanisms with yaw moment. The latter method provides for a rather sophisticated control system and expensive sensors. However, there is the third approach which implies using neural networks to control the speed of electric vehicles. The method is simple, universal, and very flexible. However, most studies into this issue do not provide methods for developing this system in real form and present only theoretical calculations. This paper and research aim to provide materials and software to realize electronic differential based on artificial neural networks into practice.