Title: Vehicle lane change driving intent recognition based on Bayes-XGBoost

Authors: Qinghui Zhou; Shunjie Sun

Addresses: School of Mechanical-Electronic and Vehicle Engineering, Beijing University of Civil Engineering and Architecture, Beijing, 102627, China ' School of Mechanical-Electronic and Vehicle Engineering, Beijing University of Civil Engineering and Architecture, Beijing, 102627, China

Abstract: Accurate identification of lane-changing intentions is critical for ensuring the safety of autonomous vehicles on the road. To tackle this challenge, we propose a novel approach that leverages the Bayes-XGBoost model. This model integrates Bayesian optimisation techniques with the robust XGBoost algorithm, facilitating the automatic optimisation of hyperparameters. Our experiments, conducted using the NGSIM dataset, demonstrate that the Bayes-XGBoost model outperforms traditional methods, including XGBoost, long short-term memory (LSTM), support vector machine (SVM), GraphSAGE networks, and graph convolution network (GCN), in accurately identifying lane-changing intentions. The primary innovation of the Bayes-XGBoost model lies in its ability to effectively capture complex traffic patterns and dynamically adjust model parameters for optimal performance. By accurately predicting lane-changing intentions, autonomous vehicles equipped with this model can make informed decisions to prevent potential collisions and ensure safe driving behaviours. Therefore, the Bayes-XGBoost model holds considerable promise for enhancing traffic safety by mitigating accident risks and improving overall road safety.

Keywords: intelligent vehicles; vehicle lane change; XGBoost; Bayesian optimisation; lane change intent recognition.

DOI: 10.1504/IJVD.2024.146773

International Journal of Vehicle Design, 2024 Vol.96 No.3/4, pp.243 - 262

Received: 26 Mar 2024
Accepted: 15 Nov 2024
Published online: 17 Jun 2025 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article