Vehicle Fuel Consumption Prediction Based on Least Squares Support Vector Machine Optimized by Improved Particle Swarm Optimization Algorithm

Abstract

Firstly, this paper summarizes the characteristics of vehicle running characteristics and design parameters, which have influence on vehicle fuel consumption. Secondly, 200 vehicle’s test results are used as training samples, with sensitive features and fuel consumption of type approve test as the input parameters, and the actual vehicle fuel consumption as output parameters. Then, a vehicle fuel consumption prediction model, which is based on least squares support vector machine, is established. Finally, the vehicle fuel consumption prediction model is used to predict the fuel consumption of another 100 vehicles. The results show that the prediction error of the test samples are less than 5%, and the fuel consumption prediction model proposed in this paper has fully considered the impact of vehicle operating characteristics and design parameters on fuel consumption. In addition, the fuel consumption prediction model has high prediction accuracy and reliability than some traditional methods such as backpropagation neural network (BPNN).