Real-time Prediction of Autonomous Driving Braking Distance Based on LSTM - Kinematics Model

With the rapid development of autonomous driving technology, real-time prediction of braking distance is of vital importance for ensuring driving safety. This paper proposes an LSTM - kinematics model fusion method, which combines the Long Short-Term Memory network (LSTM) with the vehicle kinematics model to achieve high-precision prediction of braking distance in complex environments. Through feature extraction and time series analysis of multi-source data such as vehicle speed and road conditions, LSTM is used to dynamically capture nonlinear factors, while the kinematics model provides physical constraints. Experiments were conducted on datasets including different road conditions such as dry, slippery, and snowy roads. The results show that the average absolute error (MAE) of this model is <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{1. 2 3}$</tex> meters, which is <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{4 2. 7 \%}$</tex> lower than that of the traditional kinematics model. The single-step prediction time is only 18.6 ms, meeting the real-time requirements of the autonomous driving system. The research results provide an effective solution for improving the safety of autonomous driving.