Lane change prediction for autonomous vehicles using trajectory templates

Date

2024-8-26

Author

Yıldız, İsmail Hakkı

Metadata

Show full item record

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Item Usage Stats

74
views

0
downloads

Cite This

The navigation of autonomous vehicles depends both on the desired motion of the ego-vehicle and on the motion of surrounding traffic such as vehicles. Accordingly, the fast and accurate trajectory prediction of vehicles is crucial for safe and efficient autonomous navigation. This thesis introduces a novel template-based approach for lane change trajectory prediction. The method utilizes clothoid-based bi-elementary paths, a compact and flexible mathematical model, to represent template lane change trajectories. The parameters of these templates are then estimated using least-squares fitting of measured vehicle motion data. In our method, we employ an efficient approximation of bi-elementary paths by arc-splines, that is, sequences of arc segments. The main advantage of this approximation is the representation of lane change trajectories by a very small number of parameters, which reduces the computational complexity of curve fitting. An additional advantage is the direct applicability of our method to lane changes on both straight and curved roads through appropriate coordinate transformations. A comprehensive evaluation on synthetic data as well as the real-world NGSIM dataset, using metrics like Root Mean Squared Error (RMSE) in spatial and temporal coordinates, demonstrates the method's competitive accuracy and superior computational efficiency compared to state-of-the-art methods. Our approach is further characterized by its interpretability and real-time performance and, hence, offers a promising solution for enhancing the safety and decision-making capabilities of autonomous vehicles.

Subject Keywords

Lane change prediction, Trajectory template, Autonomous vehicles, Clothoid, Curve fitting

URI

https://hdl.handle.net/11511/111408

Collections

Graduate School of Natural and Applied Sciences, Thesis