Anteumbler: Non-Invasive Antenna Orientation Error Measurement for WiFi APs

Dawei Yan, Panlong Yang, Fei Shang, Nikolaos M. Freris, Yubo Yan

Introduction

Indoor localization systems such as SpotFi and LocAP have reached high levels of precision by accurately mapping the locations of WiFi Access Points (APs). However, a critical factor is often overlooked: the 3D orientation of the antennas. In real-world deployments, antennas are rarely perfectly vertical due to haphazard installation, physical bumps, or intentional tilting for coverage.

The Impact of Orientation Error

Research into infrastructure configuration reveals that even minor orientation discrepancies can significantly degrade localization accuracy. As shown in the figure below, an orientation error of just 8° can result in a median localization error of 1 meter in a standard 3000 sq. ft. office environment. This error is comparable to a 30cm drift in the physical location of the AP itself.

Impact of orientation error on localization accuracy.

Anteumbler: A Non-Invasive Solution

Anteumbler is a novel system designed to measure the 3D orientation (yaw, roll, pitch, and irregular tilt) of WiFi antennas using standard WiFi signals. Unlike existing methods that require manual inspection, specialized sensors, or line-of-sight computer vision, Anteumbler is non-invasive and works with off-the-shelf hardware.

Technical Methodology

The system is built upon the Spatial Angle Model, which utilizes the Friis transmission formula. It identifies that received signal power is maximized when the transmitter and receiver antennas are perfectly aligned in directivity and polarization. The workflow consists of three primary stages:

  1. Removing Unknown Parameters: Using the Spatial Angle Model to isolate orientation from hardware-specific variables.
  2. Iterative Optimization: Instead of a time-consuming brute-force search, Anteumbler employs an iterative algorithm that reduces measurement time from hours to just 135 seconds per AP.
  3. Dual Plane Intersection: By intersecting two geometric planes, the system calculates orientation without requiring the exact distance between the measurement device and the AP.
The Anteumbler workflow for estimating antenna orientation.

Experimental Evaluation

The system was evaluated using a TurtleBot equipped with an Intel IWL5300 NIC to collect Channel State Information (CSI). Testing took place in two distinct environments:

  • The Hall: A 6,500 sq. ft. area with weak multipath interference.
  • The Office: An 1,800 sq. ft. area with strong multipath and Non-Line-of-Sight (NLoS) conditions created by wooden boards and pillars.

Key Results

  • Precision: Anteumbler achieved a median measurement error of less than 6° for both elevation and azimuth angles across various antenna types.
  • Localization Improvement: Integrating Anteumbler with the SpotFi localization system reduced the median user localization error by more than 1 meter.
  • Robustness: The system maintained high accuracy (median error < 8°) even in dynamic environments with human movement and complex multipath scenarios.

Conclusion

Anteumbler demonstrates that antenna orientation is a vital yet manageable component of indoor localization. By providing a fast, non-invasive way to calibrate infrastructure, it paves the way for self-correcting WiFi networks that maintain high precision over time without manual intervention.

Summary updated

BibTeX
@inproceedings{yan2024anteumbler,
  author = {Dawei Yan and Panlong Yang and Fei Shang and Nikolaos M. Freris and Yubo Yan},
  title = {{Anteumbler: Non-Invasive Antenna Orientation Error Measurement for WiFi APs}},
  booktitle = {IWQoS},
  year = {2024},
  doi = {10.1109/iwqos61813.2024.10682937},
}
Figure