Technical training software platform using machine learning and AR for Air Force maintenance and service operation

A comprehensive technical training software platform combining computer vision and Augmented Reality (AR) technologies can transform the human learning process and experience. Computer vision is leveraged to understand the physical environment, while AR is used to superimpose contextual information onto the physical world. We aim to provide software solution using AR tools and systems to address specific challenges in technical training of physical systems. One challenge with existing AR software platforms is failure of recognizing complex systems or objects that have small, shiny, or textureless components. In this STTR Phase I proposal, we propose to develop specific machine learning algorithms that can be integrated with state-of-the-art AR technology, to enhance automatic object and component recognition and display automatic annotations via AR. With the capability of recognizing components more reliable, accurate, cheaper and faster, the developed machine learning algorithms can improve technical training efficiency on complex physical systems using existing AR application platform. Focus area: FA-002 Artificial Intelligence (AI).