Analysis and Accuracy Improvement of UWB-TDoA-Based Indoor Positioning System

Paolo Grasso, Mauro Sebastián Innocente, Jun Jet Tai, Olivier Haas, Arash Moradinegade DizqahSensors 2022 · DOI

TL;DR

This paper tackles the challenge of improving the accuracy and reliability of Ultra-Wideband (UWB) Time Difference of Arrival (TDoA) based indoor positioning systems. By analyzing precision limits, identifying failure zones, and developing a novel debiasing filter, the authors demonstrate significant accuracy improvements, making UWB positioning more robust for critical applications.

What Problem Does This Paper Address?

Accurate and reliable indoor positioning is crucial for a wide range of applications, from asset tracking and robotics to assisted navigation. While Ultra-Wideband (UWB) technology offers high precision and data rates for indoor environments, its performance can be hampered by signal disruptions and inherent limitations in positioning algorithms. Specifically, existing systems often struggle with:

  • Precision Variability: The accuracy of UWB positioning isn't uniform across the entire indoor space.
  • Failure Zones: Certain regions, particularly near anchors or due to geometrical constraints, can lead to complete system failure or unreliable estimates.
  • Systematic Bias: Even when operational, UWB systems can exhibit systematic biases that reduce overall accuracy.

This paper aims to address these issues by providing a comprehensive analysis of UWB TDoA system performance, including its precision limits and failure modes, and by proposing a practical solution to mitigate systematic biases and enhance accuracy.

Methods

The authors employ a multi-faceted approach to analyze and improve the UWB TDoA positioning system:

  • Cramér-Rao Lower Bound (CRLB) Analysis: This theoretical framework is used to determine the fundamental limits of precision for the UWB TDoA system. Crucially, the analysis incorporates the anisotropic radiation pattern of the UWB antennas, moving beyond idealized isotropic assumptions.
  • Bifurcation Envelope Analysis: To identify regions where the system is prone to failure, the authors utilize a geometrical study based on bifurcation curves. This analysis defines "bifurcation envelopes" that bound these unreliable areas.
  • Flyable Area Definition: By combining the CRLB analysis (for precision) and the bifurcation envelope analysis (for failure zones), the paper defines a "flyable area" – the region where the system is both precise and reliable.
  • Extended Kalman Filter (EKF): The system under study already incorporates an EKF for initial state estimation.
  • Debiasing Filter (DF): A novel debiasing filter is developed to address the systematic bias observed in the EKF's position estimates. This filter works by:
  • Building a bias map based on measured positions rather than actual positions.
  • Using Radial Basis Function Networks (RBFN) for interpolation to create continuous bias surfaces.
  • Subtracting the estimated bias from the measured positions to obtain a debiased estimate.

Datasets

Not discussed in the paper. The experiments were conducted using a custom setup with bitcraze's Loco Positioning System and a Crazyflie 2.0 drone.

Key Insights and Ideas

  • Anisotropy Matters: The anisotropic radiation pattern of UWB antennas significantly impacts precision, leading to ripples in the expected precision contours. This highlights the importance of considering real-world antenna characteristics.
  • Flyable Area is Key: Defining a "flyable area" by combining precision bounds (CRLB) and failure zones (bifurcation envelopes) provides a more realistic assessment of system usability than precision alone.
  • Bias is Position-Dependent and Measurable: The study confirms that biases are not uniform and can be mapped as a function of measured positions, making them amenable to correction.
  • Debiasing Filter Effectiveness: The proposed debiasing filter significantly improves accuracy by effectively compensating for systematic biases, even in dynamic scenarios.

Key Results

  • Precision: The best precision is achieved within the convex hull of the anchors, estimated to be around ±3 cm.
  • Failure Zones: The system fails near the anchors, within the defined bifurcation envelopes.
  • Accuracy Improvement (Static): The debiasing filter improved accuracy by approximately 15 cm on average for static measurements.
  • Accuracy Improvement (Dynamic): For dynamic measurements, the debiasing filter improved accuracy by about 5 cm on average.
  • Flyable Area: The paper successfully defines a flyable area where positioning is reliable.

Limitations and Future Work

  • 2D Focus: The current study is limited to 2D indoor positioning. Future work could involve generalizing the approach to 3D IPSs.
  • Anchor Placement: While the study considers anisotropic antenna patterns, further optimization of anchor number, positioning, and orientation in 3D could be explored.
  • Filter Refinement: The debiasing filter's performance can be affected by sampling density in the bias map. A finer sampling mesh could potentially improve accuracy in certain areas. The authors also suggest that other filtering techniques might be needed to address dynamic misbehaviors not handled by the current DF.
  • Automation: The process of defining the flyable area and building bias maps could be automated for easier application in different settings.

Summary updated

BibTeX
@article{grasso2022analysis,
  author = {Paolo Grasso and Mauro Sebastián Innocente and Jun Jet Tai and Olivier Haas and Arash Moradinegade Dizqah},
  title = {{Analysis and Accuracy Improvement of UWB-TDoA-Based Indoor Positioning System}},
  journal = {Sensors},
  year = {2022},
  doi = {10.3390/s22239136},
}
Figure