Previous Abstract Return to Session E9 Next Abstract

Session E9: PNT for Autonomous Systems

Ultra Low SWAP-C Vision-Based Navigation for Small UAS
Eric Farmer, Jeremiah Ritchey, Frank Cavallaro, and Kevin Betts, Leidos, Inc.
Location: Ballroom B
Date/Time: Wednesday, Jun. 3, 9:15 a.m.

Low-cost small unmanned aerial systems (sUAS) need the ability to navigate accurately when GPS is compromised, but current alternative navigation technologies are costly compared to the platform cost, especially for attritable assets. By leveraging the mass-production capabilities of consumer off-the-shelf (COTS) vendors, Leidos brought low-cost vision-based navigation (VBN) to sUAS platforms in two different solutions. In the first solution using the Leidos Mobile VBN (MoVBN), Leidos achieved 18m RMSE 3D GPS-denied performance on multicopter Group 1 sUAS. In the second solution using a custom designed-for-manufacture prototype unit, Leidos VBN achieved 11m RMSE 3D fully RF-denied performance for 60 minutes of flight at Eglin AFB on a fixed-wing Group 3 UAS. Both solutions are designed to be bolt-on to sUAS for rapid integration and have been tested with unmodified open source Ardupilot flight controllers. These results demonstrate Leidos VBN is feasible on low-cost COTS hardware.

Background and Objectives: Leidos has a long legacy providing mature navigation solutions for aerial vehicles in GPS-compromised environments. The core of these solutions is built on Leidos’ in-house georegistration map matching for absolute position measurements, visual-inertial odometry for relative speed estimation, and a dynamically reconfigurable particle filter for fusing all measurements into a current position velocity attitude estimate. Additionally, the Leidos framework is built on the All Source Position and Navigation (ASPN) standard, which allows easy integration with other sensors and vendors.
The objective was to leverage Leidos’ existing software solutions and deploy them on ultra low-cost consumer off-the-shelf (COTS) hardware to create a bolt-on navigator for sUAS to support GPS-compromised missions.

Methodology and Key Innovations. In the development of a low-cost solution, two approaches were pursued. The first approach identified a COTS hardware package that included the necessary sensing suite for VBN, which allows for mass-production cost savings. The second approach built a custom unit using COTS components, which allows for better navigation sensors to be integrated for improved accuracy.

Results, Conclusions, and Significance. Leidos MoVBN was tested at multiple events using multicopter Group 1 sUAS operating between 60m to 165m above ground level (AGL) and achieved 18m RMSE 3D accuracy. These outputs have been successfully demonstrated running closed-loop with sUAS platforms running standard commercial autopilots.
Leidos VBN on the custom unit was demonstrated at an AFWERX event at Eglin AFB in August 2025 where it was mounted to a fixed-wing Group 3 UAS operating at 165m AGL. With 60 consecutive minutes of RF-denial, Leidos VBN achieved 11m RMSE 3D accuracy.
This presentation will provide details on the integration and performance of low-cost VBN with sUAS. The primary conclusion is that VBN is available for Group 1 and Group 2 sUAS. We will discuss at the CUI level which hardware components were selected for these platforms based on the unique requirements of sUAS size, weight, power, and cost constraints.

Previous Abstract Return to Session E9 Next Abstract