7 - Artificial Intelligence Capabilities for Aviation Behavior Characterization & Anomaly Detection

ATAC leverages our 20+ years of experience in aviation surveillance data analytics and modeling to develop significant capabilities for Navy SBIR Subtopic N193-A001 focus area 7. Our approach applies Machine Learning (ML) and Artificial Intelligence (AI) techniques to ADS-B data for behavior characterization (BC) and anomaly detection (AD). The proposed approach is innovative because: (1) It applies ATAC’s validated trajectory analysis algorithms to create richer datasets for training AI algorithms, (2) It leverages ATAC’s expertise with NASA’s proven Multiple Kernel Anomaly Detection (MKAD) algorithm to provide a low risk approach for BC & AD on ADS-B data. MKAD has been successfully demonstrated on radar flight tracks before, (3) It applies ATC SME feedback to re-training the AI anomaly detection algorithms, and (4) It provides an innovative deep learning approach as a backup. It addresses the Navy requirement of creating and evaluating AL/ML algorithms to: (1) identify apparent air corridors and (2) detect anomalous behavior in support of determining aircraft intent, using ADS-B data. The technical objectives include: (1) Determining the overall system and functional requirements for the AI BC & AD system, (2) Developing prototype BC & AD components, and (3) Demonstrating the feasibility of the BC & AD approach.