A Low-Cost Collaborative Location Scheme with GNSS and RFID for the Internet of Things

Changfeng Jing, Shouqing Wang, Mingshu Wang, Mingyi Du, Lei Zhou, Tiancheng Sun, Jian WangISPRS Int. J. Geo Inf 2018 · DOI

TL;DR

This paper introduces a novel, low-cost system for locating public infrastructure assets in smart cities by combining Global Navigation Satellite System (GNSS) and Radio Frequency Identification (RFID). By using a single mobile RFID reader mounted on a GNSS receiver, the system eliminates the need for numerous expensive reference tags, significantly reducing costs and simplifying deployment, while achieving sufficient accuracy for asset management.

What Problem Does This Paper Address?

The proliferation of smart cities and the Internet of Things (IoT) necessitates efficient ways to manage and inspect vast numbers of public infrastructure assets. Accurately locating these assets is crucial for real-time monitoring, maintenance, and emergency response. Existing RFID-based localization systems often suffer from high deployment costs (requiring many RFID readers or reference tags) and complex installation, making them impractical for large-scale applications, especially in challenging environments or during emergencies. GNSS, while a common outdoor localization method, has limitations in urban canyons and indoor/underground areas due to weak signals. This paper aims to bridge this gap by proposing a system that is both cost-effective and easily deployable for accurate infrastructure localization.

Methods

The core of the proposed system is a collaborative approach that leverages both GNSS and RFID. Instead of relying on a dense network of fixed reference tags, the system uses a single mobile RFID reader equipped with a GNSS receiver. This mobile unit acts as a dynamic reference point.

Here's a breakdown of the methodology:

  • Hardware Setup: A mobile RFID reader is integrated with a GNSS receiver. The target infrastructure asset is fitted with an RFID tag.
  • Data Acquisition: As the mobile reader moves, it records its GNSS-measured position ($p_i$) and the distance ($d_i$) to the RFID tag. This generates a series of tuples $(p_i, d_i)$.
  • Localization Principle: The system employs a triangulation-based approach. With at least three such tuples, the system can form triangles where the reader's positions and the measured distances define circles. The intersection of these circles (or more accurately, the solution to a system of equations derived from these circles) provides candidate locations for the RFID tag.
  • Addressing Ambiguity: The algebraic solution to the equations can yield two possible tag locations. The paper addresses this by using a "redundant observation methodology," where more than the minimum required measurements are taken. This results in multiple candidate points for the tag's location.
  • Distance-Based Weighting Algorithm: To improve accuracy and account for measurement errors, the paper proposes a distance-based proportionality function. Instead of a simple average of candidate points, a power-law based weight is applied, giving more importance to closer measurements and less to those potentially affected by greater signal noise. The weight $w_i$ is calculated as $w_i = \frac{1}{(d_i)^\lambda}$, where $\lambda$ is an optimized parameter.
  • Linear Interpolation for Synchronization: A challenge is the differing signal frequencies of GNSS and RFID. To overcome potential time synchronization issues, a linear interpolation algorithm is used to estimate GNSS positions at the precise times when RFID distance measurements are taken.

Datasets

The paper does not explicitly mention using pre-existing datasets for training or evaluation. Instead, it describes an experimental setup conducted on the university campus of Beijing University of Civil Engineering and Architecture, China. The "ground truth" coordinates for the target object were measured using a Leica TS06 Total Station. The experimental data consists of the trajectories of the mobile RFID reader and the corresponding RFID tag distance measurements.

Key Insights and Ideas

  • Cost Reduction through Mobility: The most significant insight is the clever use of a mobile RFID reader with GNSS to replace a large number of static reference tags. This dramatically cuts down on hardware costs and installation complexity, making large-scale deployment feasible.
  • Dynamic Reference Points: The system treats the moving GNSS-equipped RFID reader as a dynamic reference, allowing for flexible and rapid deployment without pre-surveying fixed tag locations.
  • Improved Accuracy via Weighted Averaging: Moving beyond simple averaging of candidate points, the distance-based weighting algorithm is a crucial refinement. It acknowledges that signal quality and noise are distance-dependent, leading to more robust and accurate final estimations.
  • Practicality for Urban Environments: The proposed scheme is designed to be practical for smart city applications, where infrastructure is often dense and varied, and rapid deployment might be critical (e.g., during emergencies).

Key Results

The experiments demonstrated the effectiveness of the proposed scheme:

  • Accuracy Improvement: The average error achieved by the proposed system was 1.3904 m. This is a significant improvement compared to using only a GNSS receiver outdoors, which resulted in an error of 14.1497 m.
  • Low-Cost and Easy Deployment: The system requires only one mobile RFID reader and one portable GNSS receiver, making it significantly cheaper and easier to deploy than systems requiring numerous reference tags.
  • Parameter Optimization: Sensitivity analysis showed that the parameter $\lambda$ in the weighting algorithm achieved the best results when set to 3.
  • Comparison of Algorithms: The distance-weighted mean algorithm improved the average accuracy from 2.5095 m to 1.3904 m compared to the algebraic mean algorithm.

Limitations and Future Work

  • Hardware Synchronization: The paper acknowledges that synchronizing the RFID signal and the GNSS signal relies on a computer, which can be a challenge for achieving higher accuracy. Future work could involve hardware-level signal synchronization.
  • Outlier Detection: The authors suggest that outlier detection algorithms could further optimize the RFID and GNSS signals, leading to improved location accuracy.
  • Further Applications: The researchers plan to expand the application of their location systems to location-based business applications and services in smart city initiatives.
  • GNSS Signal Weakness: While the system aims to overcome GNSS limitations, it's still dependent on GNSS signals. In extremely challenging environments (deep indoors, tunnels), its performance might be limited, though RFID can still provide proximity information.

Summary updated

BibTeX
@article{jing2018low,
  author = {Changfeng Jing and Shouqing Wang and Mingshu Wang and Mingyi Du and Lei Zhou and Tiancheng Sun and Jian Wang},
  title = {{A Low-Cost Collaborative Location Scheme with GNSS and RFID for the Internet of Things}},
  journal = {ISPRS Int. J. Geo Inf.},
  year = {2018},
  doi = {10.3390/ijgi7050180},
}
Figure