Optimized Simultaneous Aided Target Detection and Imagery based Navigation in GPS-Denied Environments

Introduction

Unmanned Aerial Vehicles (UAVs) make significant contributions to the warfighting capability of operating forces. Reconnaissance, intelligence, surveillance, and target detection, recognition and tracking are premier missions for UAVs. To extend the scope of operation of UAVs to support these missions, their safe navigation despite an intermittent or lacking GPS signal must be assured. Current and future manned and unmanned aerial vehicles require both navigation and Aided Target Detection/Aided Target Recognition (AiTD/AiTR) capabilities in Anti-Access Area Denial (A2AD) environments. Combining both these functions into a compact, lightweight processing form factor suitable for UAVS is an important but challenging task.

In this presentation, we describe and demonstrate a comprehensive optimized vision-based real-time solution to provide Simultaneous Aided Target detection and Imagery based Navigation (SATIN) capabilities for current and future UAS in GPS-denied environments. Our solution optimizes both vision-based navigation and AiTD/AiTR into a single system, by developing and utilizing new artificial intelligence (AI) based semantic scene understanding technologies to extract visual information on automatically-identified regions of interest (ROI) perceived from an EO/IR gimbaled payload. It provides more accurate and efficient SATIN over different heights and terrains, by prioritizing the processing on semantically-informed regions across video frames. This approach also provides UAS operators with Situational Awareness of the ongoing mission.