Personalized lane departure warning based on non-stationary crossformer and kernel density estimation
Yin, H., Yue, L., Gong, Y., Li, P., & Huang, Y. (2024).
Alexandria Engineering Journal, 109, 856-870.
Summary
This study combines a Non-stationary Crossformer for departure-trajectory prediction with kernel density estimation for driver-specific warning thresholds. The algorithm is evaluated using naturalistic driving data from ten drivers in Shanghai. The study reports fewer false warnings than the comparison methods, supporting the use of personalized thresholds in lane departure warning systems.
Brief summary of the paper; see the publication record for the full abstract.