Analysts are faced with significant increases in the amount and complexity of data as sensor and imaging technologies decrease in size and expense, increase in capability, and are employed on an increasingly broad range of platforms. Managing the amount and technical complexity of data is a challenge for even trained and experienced crews. In addition, large numbers of small Unmanned Aerial Systems (sUAS), and associated sensors, are being rapidly developed and deployed, further increasing the amount of data that needs exploited. Skyward proposes to develop a novel capability to determine, assess, categorize, and counter ground threats via anomaly detection which will reduce data processing times and increase analytical focus. The Anomaly Detection Onboard Identification System (ADONIS) will implement uniquely developed algorithms and miniaturized hardware to quickly and autonomously process real-time sensor data onboard UAS platforms. Processed data will result in tips and cues to be disseminated from aircraft to the ground. These tips and cues will indicate the presence of potential anomalies within large volumes of data, so analysts can focus on the most actionable information first. Our solution builds on several proven commercial technologies and is adaptable to most DoD and commercial sUAS-based sensors (e.g., LIDAR, SAR, EO/IR, Hyperspectral). Artificial Intelligence (AI) or statistical averaging algorithms will be developed and demonstrated using existing government/commercial software and deployed to lightweight, commercial AI acceleration hardware for use on sUAS platforms. ADONIS will demonstrate the capability to more rapidly and efficiently assess threats, thereby increasing resiliency and security for Future Bases, Installations and Airmen, and substantially reducing O&M costs.
