Deep learning-based Ipoh driving cycle prediction

Abstract

The driving cycle is a series of driving behaviours, such as acceleration, braking, and cruising, that occur over a set length of time. Predicting the driving cycle can help to improve vehicle performance or anticipate the range of an electric car. Based on prior data, long short-term memory (LSTM) networks can be used to forecast a vehicle's driving cycle. This paper studies a driving cycle prediction based on LSTM by recurrent neural network (RNN) using developed driving cycle data. The objectives of this paper are; to develop an Ipoh driving cycle (IDC), to develop a prediction of future IDC and to analyze the prediction of IDC. Firstly, the driving data is collected in three different routes in Ipoh city at back-from-work times. Then the data is divided into micro-trips and the driving features are extracted. The features are used to develop a driving cycle using k-means clustering approach. The prediction is developed after the training of neural networks by using LSTM network approach with root mean square error (RSME) of 6.2252%.