Previous Abstract Return to Session C6 Next Abstract

Session C6: Terrestrial Signals of Opportunity-Based Navigation Systems

5G-Based High Precision Real-Time Location System
Eapen Kuruvilla, Anand Seshadri, Ryan Plyler, Muhammad Hussain, Denise Masi, and Mohammed Zaatari, Noblis, Inc.

Objectives
For national security, warfighting, and across the government and critical infrastructures, high-accuracy, seamless, and jam-resistant positioning systems for vehicles such as a swarm of drones are needed. No single technology can work seamlessly indoors and outdoors while providing high precision. Access to GPS can be disrupted, degraded, and/or denied. Use of complementary signals can improve both accuracy and resilience.
The goal of this real-time location system (RTLS) project is to build a high-accuracy positioning system by intelligently combining technologies such as 5G, ultrawideband (UWB), GPS, and inertial navigation system (INS) using a very fast algorithm. The sensor fusion of multiple technologies improves accuracy and reliability and thus allows seamless positioning, even if one or more of these technologies are unavailable. This algorithm can also be integrated with computer-vision based optical flow algorithms to further improve accuracy.
Methodology
Our technical approach incorporates the following design constructs:
• Vehicles receive GNSS/GPS positioning data from satellites. The reception can be interrupted due to obstructions or jamming.
• Vehicles receive 5G positioning data when cellular coverage is available, and the feature is enabled by the service providers. 5G positioning is expected to be available in 2025; until then we are using 4G positioning. Cellular connectivity is also used for communication with an edge server.
• UWB is primarily used for measuring the distance, or ranging, between two vehicles. The sensor fusion algorithm uses the UWB range measurements to improve the accuracy of the absolute positioning offered by GPS and 5G.
• INS is used to estimate the position relative to a reference location. The INS is subject to drift errors, hence periodical corrections based on 5G or GPS positioning are required.
• The sensor fusion algorithm combines the GPS, 5G, INS, and UWB positioning data; each measurement is given a weight proportional to its estimated accuracy when fusing them. The algorithm operates in a fully distributed mode, and each vehicle receives the estimated locations of its neighbor vehicles via UWB data links. After performing the sensor fusion locally, the vehicle shares its estimated location with its neighbors via UWB data links.
As an additional branch of research, we explored the use of computer vision-based algorithms and concepts as a primary and secondary method of navigation.
Results
We demonstrated our algorithm through both simulation and a real hardware prototype, achieving significant improvements in positioning accuracy. Key results are as follows:
• Sensor Fusion Improvement: Our approach improves positioning accuracy by 46% compared to the existing GPS plus INS configuration for navigation.
• INS Optimization: INS now leverages corrected positions from previous measurements, requiring about six location measurements for convergence.
• Our algorithm converges 50 times faster than the standard optimizations methods such as ordinary least squares (OLS).
Further enhancements focus on providing resilience in GPS-denied environments:
• Enhanced Fusion Algorithm: By incorporating temporal diversity, our algorithm continues to maintain accuracy in a GPS denied environment.
• Real-World Results: As vehicles move away from GPS-covered areas, relative motion between vehicles significantly improves accuracy, utilizing the motion itself to create additional anchors in the temporal domain.
Conclusions and Significance
This project was able to address the following mission objectives:
• We developed a fast algorithm for a highly accurate and reliable positioning system by intelligently combining 5G, UWB, GPS, and INS, enabling resilient navigation and asset tracking use cases for government agencies. This multi-tech sensor fusion improves accuracy during signal jamming.
• We developed a physical hardware prototype of this system.
• We prototyped a computer-vision based location algorithm

Previous Abstract Return to Session C6 Next Abstract