Prediction of bicycle dynamics parameters based on machine learning

Abstract

There are many control methods for the stable motion of a bicycle, and some scholars have verified that the stable motion of the bicycle can be achieved through the proportional control of the inclination of the body. This paper combines machine learning with bicycle dynamics parameter prediction, and uses neural network to fit the relationship between the dynamic parameters at time t and t+Δt, and predict the inclination of the body at time t+Δt. The fitting result proves that the neural network method can effectively fit the law, and the prediction effect is better.