Picosat Pattern of Life (AFWERX Topic 15276)

The primary objective of this proposed effort is to reach technology readiness level (TRL) 8 for a novel capability to improve the battlefield situational awareness (SA) of both mounted and dismounted United States warfighters across a variety of “Converged” battlefield environments using space- and ground-based assets. The novel capability involves rapid prototyping of low-resource learning algorithms for pattern of life and anomaly detection (POL/AD) embedded in flocks of low-power, size, weight and price picosats to provide real-time SA.  In addition to intelligence, surveillance and reconnaissance (ISR) of ground-based assets and activities, the battlefield SA includes on-demand placement for ISR of other spacecraft. Picosats capable of supporting these requirements are available on the commercial market through the PocketQube product line manufactured by the Offeror. The solution supports terrestrial situational awareness with a low-size, -weight and -power and -price (SWaP2) picosatellite platform to achieve dedicated and persistent intelligence, surveillance and reconnaissance (ISR) from orbit, while leveraging novel and optimized low-resource (machine) learning (LRL) algorithms for continuous monitoring of pattern of life and anomaly detection in real time, thereby identifying the most timely, relevant and important information collected. The solution involves space payloads containing sensor systems that primarily perform earth observation in the infrared, ultraviolet and visible light spectra approaching one meters/pixel resolution by leveraging multiple satellites in a flock.  The payloads are at the picosatellite (“PocketQube”) level (1P, 2P, 3P).  These picosatellites contain custom software and algorithms that use low-resource (machine) learning (LRL) techniques to automatically identify Objects and Activities that fall outside the "normal" pattern of life observed by the picosatellite constellation in real time.  The effort and time required to train machine learning models using LRL is significantly less than other known approaches.  Pattern of life is maintained on the Objects and Activities, and anomalous Objects and Activities are relayed to Ground Stations (for reach-back support) and mobile receivers (carried by mounted and dismounted units), and are readily integrated into other data systems to improve situational awareness. This pattern of life and anomaly detection (POL/AD) capability is fully supported on Air Force classified networks in the Offeror’s commercial off-the-shelf (COTS) software product Watchman for Defense (W4D), which is readily adaptable for application on space-based assets.