US Air Force AI Initiative for Cislunar Space Vehicles

The US Air Force has launched a search for off-the-shelf artificial intelligence and machine learning software for autonomous spacecraft operating in the cislunar space domain beyond Earth orbit. According to the published request for information document, the industry is asked to submit flight-ready concepts to overcome communication delays at vast distances from Earth.

In a request for information (RFI) document, the US Air Force stated that it has a policy of developing a sustainable architecture beyond Earth orbit. In this context, flight-ready concepts capable of performing autonomous maneuvers such as station-keeping, orbit optimization, and hazard avoidance for objects moving in the cislunar domain are being explored. The deadline for responses is set for December 4, and these feedbacks are anticipated to shape future requirements.

The operation of cislunar systems in a much larger volume compared to near-Earth systems can lead to communication delays in ground command and control systems. The document emphasizes that these delays resulting from the increased volume render traditional ground-based command and control systems unsustainable within operational timelines. Therefore, the Air Force is requesting industry expertise in a white paper format under five headings: autonomous mission planning, predictive analytics, trajectory planning, computer vision, and other technical areas.

The service branch wants fully developed autonomy software that requires no development within 90 days following an Authority to Process order, can operate with minimal training, and can become fully qualified within one year. The request for information document underlines that rather than investing in a cislunar research and development program, the government wants to see the existing software capacity that the industry can deliver rapidly. It is stated that investments for potential future work may be shaped in line with the information obtained from the process.