SBIR-STTR Award

The Base of the Future is a weapon system that must be both resilient and effective in the delivery of its mission. As a part of mission readiness, base perimeter defense is essential to safeguard operations, personnel, and assets from directed or inadvertent disruptions perpetrated by foreign/unauthorized operators.  Having a wide range of security concerns from an inebriated individual wondering into base property or breaking through fencing to specific threat actors seeking to disrupt or document base activities, monitoring large areas in challenging terrain is no easy task. Additionally, over reliance in local authorities for notification of occurrences surrounding the base can lead to “embarrassing” situations. The work proposed in this SBIR application addresses base security needs specifically in response to the re-construction efforts at Tyndall Air Force Base, as discussed at the AFWERX Challenges entitled “Base Perimeter and Defense” and “Leveraging Operational Technology”. In light of Airgility’s participation at the AFWERX Challenge Workshops, to address Tyndall’s coast line security needs, Airgility Inc. proposes to develop a low cost object tracking turret-like solution whose slew-to-cue capability is autonomously controlled by a cooperating stack of light-weight computational algorithms for minimal tracking latency using vision-based hardware. Each object tracking camera node is entirely self-contained (independent) and has its own trained data (ML) implementation to aid in localized decision-making such as friend-or-foe object identification and false alert rate (FAR)/nuisance alert rate (NAR) mitigation. The hardware is intended for maximum portability, minimal footprint, quick deployment/removal, unit-to-unit blanketed coverage, and target tracking hand-off capability between individual nodes.  The human behavior being replicated with the camera nodes follows a similar premise as tracking a person who has fallen overboard. In that case, a spotter both physically points towards the victim overboard while verbally directing others to gaze in the same direction. By doing this, more spotters keeping track of the victim reduces that chances of losing the victim. The camera nodes track in similar fashion, where the sight picture is analyzed via algorithms and ML, and the “finger pointing” is performed via azimuth and communicated between local nodes. However, if no azimuth is provided by a neighboring node, the camera operates on standby scanning mode. A key function of the Intelligent Littoral Awareness System is the ability to track and identify (using ML object detection at the edge; non-cloud based) intrusions/intruders transitioning from the marine domain (surf) and onto the terrestrial domain (beac