ISR-Brain Machine Intelligence for Unmanned Aircraft Systems

Abstract

This paper presents a system for extrapolating knowledge and classification rules from existing ISR FMV and creating an ISR-Brain. As combat operations have grown to depend upon assured, live ISR support during operations, US forces are presented with formidable challenges to integrate artificial intelligence (AI) capabilities with existing ISR systems. The common challenge being the variance at which advances in commercial and academic AI are deployed compared to rate of speed that innovative AI systems are developed and utilized in military domains. ISR, USAF and SOCOM need to develop a means to seamlessly integrate military and commercial state-of-the-art systems. The ISR-Brain presented will be capable of converting classifiers in existing ISR FMV to machine learning rules for real time ISR sensor, multi-source, multi-enclave data and adaptable with ongoing research efforts with A2, SOCOM, JIEDO, MITRE and Project MAVEN to develop and test and ISR-Brain to enable the system to integrate with all ISR sensors and predict future Troops in Contact events (TIC) and IED events.