Techniques to Camouflage and Deceive AI/ML or Autonomous Systems on the Battlefield

The Defense Systems Information Analysis Center (DSIAC) was asked to research techniques for defeating adversary artificial intelligence/machine learning (AI/ML) target recognition. AI/ML research is driving the advancement of the next generation of automatic target recognition (ATR) capabilities using electro-optical infrared (EO/IR) sensors. The inquirer was seeking to develop and field robust AI/ML-based ATR algorithms and, at the same time, recognized that it is critical to develop concealment and deception technologies to protect against peer and near-pear adversaries’ AI/ML ATR capabilities. DSIAC subject matter experts identified and compiled findings on a variety of options from open sources, including flooding the light source; using near-infrared light, retroreflectors, and masking; and using specialized images to confuse AI/ML, labeling, and future camouflage options.