[Paper Review] Using Ultra-Wideband Technology in Vehicles for Infrastructure-free Localization

Rusheng Zhang, Lin Song+6 more|arXiv (Cornell University)|Feb 20, 2019

Indoor and Outdoor Localization Technologies19 references4 citations

TL;DR

This paper proposes an infrastructure-free vehicle localization system using Ultra-Wideband (UWB) technology with on-vehicle UWB anchors mounted on rear-view mirrors. It demonstrates through theoretical analysis and field trials that the two-anchor configuration achieves sufficient accuracy—distinguishing pedestrian positions on either side of the vehicle up to 30m away—enabling applications like Virtual Pedestrian Traffic Light (VPTL), where pedestrian crossing detection is critical.

ABSTRACT

In this paper, we investigate using Ultra-Wideband (UWB) technology in vehicles for localization as well as other possible infrastructure-free applications. To that end, we first introduce the on-vehicle UWB anchor layout, then conduct a theoretical analysis to shed light on the capabilities and limitations of using UWB anchors with this layout. To verify the analysis conducted, extensive field trials are performed, which yield results that match the predictions of the conducted analysis. Finally, Virtual Pedestrian Traffic Light (VPTL), an infrastructure-free pedestrian traffic light system is introduced as an example application of the presented approach.

Motivation & Objective

• To develop an infrastructure-free vehicle localization system using on-vehicle UWB anchors for applications like pedestrian detection.
• To analyze the theoretical error characteristics of a two-anchor UWB layout mounted on a vehicle’s rear-view mirrors.
• To validate the theoretical analysis with real-world field trials under varying distances and lateral positions.
• To demonstrate a practical application—Virtual Pedestrian Traffic Light (VPTL)—using the proposed UWB localization framework.
• To evaluate whether the system can reliably distinguish pedestrian positions on the left vs. right side of a vehicle in real-time.

Proposed method

• Deploying UWB devices (DecaWave DW1000) on a vehicle’s rear-view mirrors as anchors, using Double-Sided Two-Way Ranging (DS-TWR) for distance measurement.
• Using trilateration with two anchors to estimate the 2D position of a pedestrian tag (mobile UWB device), with known anchor coordinates.
• Applying a theoretical error analysis to model horizontal localization error based on anchor geometry and signal propagation.
• Conducting field experiments with 200 measurements at 10m, 20m, and 30m distances, recording lateral positions (left/right side) of the tag.
• Using scatter plots and classification techniques (e.g., SVM or hypothesis testing) to assess separation between left and right side pedestrian positions.
• Integrating the UWB localization system into the VTL (Vehicle-to-Vehicle Traffic Light) protocol to enable pedestrian phase control via VPTL.

Experimental results

Research questions

• RQ1 Can a two-anchor UWB configuration mounted on a vehicle’s rear-view mirrors achieve sufficient localization accuracy for pedestrian detection applications?
• RQ2 What are the theoretical and empirical error characteristics of this on-vehicle anchor layout in terms of horizontal positioning error?
• RQ3 Can the system reliably distinguish between pedestrians on the left and right sides of a vehicle at distances up to 30m?
• RQ4 How does the proposed infrastructure-free UWB system perform in real-world field conditions compared to theoretical predictions?
• RQ5 Can the UWB localization system be effectively integrated into existing V2V traffic management protocols to enable pedestrian phase control?

Key findings

• The on-vehicle two-anchor UWB configuration achieves a horizontal localization error of up to 1 meter, which is acceptable for non-critical infrastructure-free applications.
• Field trials confirmed that pedestrian positions on the left and right sides of the vehicle are clearly separated in scatter plots at 10m, 20m, and 30m distances.
• The system successfully distinguishes lateral pedestrian positions using simple classification methods, indicating feasibility for pedestrian crossing detection.
• The theoretical error model aligns well with empirical results, validating the analytical framework for anchor layout performance.
• The UWB-based system enables the implementation of the Virtual Pedestrian Traffic Light (VPTL) application, where pedestrian presence and crossing status can be tracked without roadside infrastructure.
• Future work may further reduce error using Kalman filtering, though current accuracy is sufficient for VPTL and similar applications.