A new approach to lane guidance systems

Eidehall, A.; Pohl, J.; Gustafsson, F. · 2005 · OpenAlex-citations

DOI: 10.1109/itsc.2005.1520145

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This paper introduces Emergency Lane Assist (ELA), a novel automotive safety function designed to address the limitations of conventional lane guidance systems, specifically false alarms during intentional lane changes and potential misuse as an autopilot. Unlike standard systems that continuously prevent lane departure, ELA integrates a threat assessment module that activates interventions only when a lane departure is deemed dangerous. The system aims to function as a pure safety mechanism rather than a convenience feature, intervening via steering wheel torque only when adjacent lanes contain approaching vehicles or other threats, while allowing safe, intentional lane changes. The core of the ELA system relies on a sophisticated tracking and estimation framework. To assess risk, the system requires detailed information about vehicle surroundings, including the positions and motion of other vehicles, as well as road geometry parameters like lane width and curvature. The authors employ an Extended Kalman Filter (EKF) to estimate these parameters. Performance is enhanced by utilizing a non-linear model within a road-aligned, curved coordinate system. This coordinate system simplifies the dynamic motion model for tracked objects by assuming they follow their respective lanes, thereby improving the accuracy of position predictions relative to the host vehicle. The decision algorithm calculates the time to cross lane markings and predicts the position of threat vehicles during this interval. If a predicted trajectory intersects a defined safety region in the adjacent lane, a collision risk is flagged, triggering the intervention. Crucially, the system also checks for threats in the host vehicle’s current lane to ensure it does not hinder evasive maneuvers. The ELA algorithm was evaluated using a demonstrator on a 300-meter straight test track with two 3.2-meter wide lanes. An inflatable dummy vehicle served as a stationary threat in the adjacent lane. Tests varied host vehicle velocity, heading angle, and lateral displacement. Results indicated that the system successfully distinguished between dangerous and safe lane changes across the tested scenarios. When a threat was detected, the lateral controller applied steering torque to prevent departure and subsequently realigned the vehicle in the center of the original lane. Driver feedback was largely positive, noting the interventions were soft and non-dramatic, providing a sense of security. However, technical limitations were observed, including vision sensor tracking loss during high-dynamic maneuvers and delayed object detection in poor visibility conditions. The study concludes that ELA offers a viable approach to reducing false alarms and misuse associated with traditional lane guidance systems. While the demonstrator proved effective in controlled scenarios, the authors note that further verification in realistic traffic environments involving moving objects, multiple vehicles, and curved roads is necessary. Future development also requires exploring additional Human-Machine Interface options beyond steering torque, such as audio or visual warnings, to enhance system usability and safety.

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.

Summary generated by qwen3.6-27b-prismaquant on 2026-06-26; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• lane positioning