Francesco De Cristofaro

arXiv2026

Lane Change Intention Prediction of two distinct Populations using a Transformer

In complex traffic scenarios, intention prediction of surrounding vehicles can improve the strategy of automated driving functions. Existing work on intention prediction is often trained on datasets of single regions or countries. In this article, a transformer network for lane change intention prediction was trained to predict whether target vehicles perform a left lane change, right lane change or keep their lane. Features inputs used were vehicle positions and distances, each for longitudinal and lateral direction. Before being inputted, these features were converted to Frenet coordinates. Two different datasets from leveLXData collected on highways were used: one from German highways and one from Hong Kong highways. Through cross-dataset evaluation, we show that the accuracy values drop to 71.94%, compared to 85.44% when doing dataset-specific training. When training on both datasets, accuracy levels up to 86.84% were achieved.

cs.LG↗

arXiv2025

Prediction of Lane Change Intentions of Human Drivers using an LSTM, a CNN and a Transformer

Lane changes of preceding vehicles have a great impact on the motion planning of automated vehicles especially in complex traffic situations. Predicting them would benefit the public in terms of safety and efficiency. While many research efforts have been made in this direction, few concentrated on predicting maneuvers within a set time interval compared to predicting at a set prediction time. In addition, there exist a lack of comparisons between different architectures to try to determine the best performing one and to assess how to correctly choose the input for such models. In this paper the structure of an LSTM, a CNN and a Transformer network are described and implemented to predict the intention of human drivers to perform a lane change. We show how the data was prepared starting from a publicly available dataset (highD), which features were used, how the networks were designed and finally we compare the results of the three networks with different configurations of input data. We found that transformer networks performed better than the other networks and was less affected by overfitting. The accuracy of the method spanned from $82.79\%$ to $96.73\%$ for different input configurations and showed overall good performances considering also precision and recall.

cs.LG↗