Vehicular Visible Light Positioning for Collision Avoidance and Platooning: A Survey

TL;DR

This survey reviews vehicular visible light positioning methods for collision avoidance and platooning, highlighting a new method that achieves cm-level accuracy and high update rates suitable for autonomous driving.

Contribution

It introduces a novel vehicular VLP method that outperforms existing techniques in accuracy and speed, advancing the state-of-the-art for vehicle localization.

Findings

The new VLP method achieves cm-level accuracy.

Simulation shows the method meets real-time localization requirements.

Challenges for real-world deployment are discussed.

Abstract

Relative vehicle positioning methods can contribute to safer and more efficient autonomous driving by enabling collision avoidance and platooning applications. For full automation, these applications require cm-level positioning accuracy and greater than 50 Hz update rate. Since sensor-based methods (e.g., LIDAR, cameras) have not been able to reliably satisfy these requirements under all conditions so far, complementary methods are sought. Recently, positioning based on visible light communication signals from vehicle head/tail LED lights (VLP) has shown significant promise as a complementary method attaining cm-level accuracy and near-kHz rate in realistic driving scenarios. Vehicular VLP methods measure relative bearing (angle) or range (distance) of transmitters (i.e., head/tail lights) based on received signals from on-board photodiodes and estimate transmitter relative positions based on those measurements. In this survey, we first review existing vehicular VLP methods and propose a new method that advances the state-of-the-art in positioning performance. Next, we analyze the theoretical and simulated performance of all methods in realistic driving scenarios under challenging noise and weather conditions, real asymmetric light beam patterns and different vehicle dimensions and light placements. Our simulation results show that the newly proposed VLP method is the overall best performer, and can indeed satisfy the accuracy and rate requirements for localization in collision avoidance and platooning applications within practical constraints. Finally, we discuss remaining open challenges that are faced for the deployment of VLP solutions in the automotive sector and further research questions.