Enabling an EO-Sensor-Based Capability to Detect and Track Multiple Moving Threats Onboard sUAS Operating in Cluttered Environments

Abstract

In an increasingly complex environment for manned and unmanned traffic, technological advancements are thought to be able to provide alerts of the incoming threats autonomously and in a timely manner. This paper explores feasibility of integration of computer vision onboard a small unmanned aerial system (sUAS) to detect and track multiple moving ground and aerial targets in the electro-optical (EO) sensor field of view. Towards this objective, the paper discusses the essence of the multiple moving targets detection (MMTD) algorithm that has been developed and tested off line already, and then proceeds with an analysis of two candidate sUAS platforms to host and run the MMTD algorithm in real time from the standpoint of hardware and software architecture design and applicability. It further continues with a description of a series of tests that were conducted progressively to assess and evaluate the overall concept with multiple sUAS and unmanned ground vehicles representing the cluttered operational environments. These tests proved the MMTD algorithm to be successful in detecting and tracking multiple moving targets, thus laying a foundation for future research on implementation of the developed techniques for autonomous vision-based collision-free operations. The paper also discusses certain hardware and software compatibility issues revealed during this study.