|Abstract:||Radio frequency interference (RFI) is a challenge that has, and continues to be, faced by the aviation industry today. This is especially true for GNSS in aviation as these signals are particularly weak. Interference comes in many forms ranging from short transient events to longer duration events, and are often caused by unintentional radiators, however all of these scenarios have the possibility to pose a risk to GNSS users. Many techniques exist to localize a signal source, however, these techniques typically rely on knowing the position from which a set of measurements are made. Therefore, when the source being localized creates a GNSS denied environment there become two problems to solve: localizing the interferer and navigating in the denied environment caused by the interferer. An unmanned aerial system (UAS) built on a multirotor platform, known as JAGER, is being developed to solve these two challenges . This paper focuses on the development of a navigation system for JAGER to enable operation in a GNSS denied environment, resulting in improvements to the RFI localization capabilities that are JAGER’s core mission. Enabling operation in the denied environment opens up new possible flight trajectories that result in significant improvements in time to localization of the RFI source. This paper presents the design, development, and flight testing of a vision based navigation system for JAGER. The resulting navigation performance is demonstrated to enable improvements in the RFI localization performance through enabling flight trajectories that take JAGER closer to the RFI source into a potentially GNSS denied environment. To quantify the improvements to the RFI localization, this paper uses the existing system’s strategy of maintaining a necessary standoff distance to an interferer to execute the mission with a GNSS position throughout the flight as a baseline. Against this baseline, this paper demonstrates the improvements possible through the use of new trajectories that bring the vehicle closer to the interferer. The navigation system is built around the use of a downward facing infrared (IR) camera and the use of optical flow to measure the velocity of the vehicle in flight. These velocity measurements are used in an extended Kalman filter (EKF) that estimates the 2D position and 2D velocity of the vehicle throughout the flight. Flight tests of the system demonstrate an ability to measure the velocity with noise of 0.7m/s resulting in a drift rate of the estimated position of 0.4% of distance traveled. This level of navigation performance enables the separation of the navigation and localization algorithms, enabling the use of two different filters for navigation and localization. A brief analysis of the effect of the noise in the velocity measurement on the navigation and localization systems is also presented. It is shown that as the noise increases above 1m/s, the localization results are no longer improved with the new trajectories taking JAGER into the GNSS denied environment. Therefore, in those cases, to combat the drift in the position estimate due to the use of velocity only measurements, this paper briefly explores the use of a simultaneous localization and mapping (SLAM) framework to use both the velocity and bearing measurements to the RFI source to simultaneously estimate the vehicle’s position and the source’s position. In simulation, this paper shows that the inclusion of bearing measurements can cap the position drift of the vehicle’s estimate and, with low noise bearing measurements, can provide improvements to the localization system in cases of high velocity measurement noise.|
Proceedings of the 31st International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS+ 2018)
September 24 - 28, 2018
Hyatt Regency Miami
|Pages:||2726 - 2736|
|Cite this article:||
Perkins, Adrien, Chen, Yu-Hsuan, Lo, Sherman, Enge, Per, "Vision Based UAS Navigation for RFI Localization," Proceedings of the 31st International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS+ 2018), Miami, Florida, September 2018, pp. 2726-2736.
ION Members/Non-Members: 1 Download Credit