Estimation of remaining driving range for a battery electric vehicle with driving cycle recognition

The accuracy of displayed remaining mileage estimation on a pure electric vehicle is very important, and the performance of accurate estimation directly affects the driver's experience. It is of great research significance to combine driving intention recognition based on test conditions, driving style, and other factors, and to estimate the remaining mileage displayed on the vehicle. This paper proposes a method for estimating remaining mileage based on probabilistic neural network (PNN) for driving intention recognition. Combined with actual road spectrum collection to extract typical working conditions, vehicle simulation results show that with driving cycle recognition, the maximum deviation between the estimated displayed mileage and the actual mileage is 4.5km, the corresponding error ratio is 1.73%, indicating that the estimated mileage is close to reality.