From NASA's rover routes to jetliners' aircraft sensors, AI is changing how the aerospace industry tests risk

In the vast expanse of space, where uncertainty reigns, NASA's Mars rover, Perseverance, navigates treacherous terrains not merely by following GPS but with the assistance of advanced AI and digital twin technology. This innovative approach allows the rover to traverse boulder fields, sand dunes, and steep slopes on the Martian surface with precision. Kevin Murphy, NASA's acting chief artificial intelligence officer, emphasizes the importance of digital twins, which serve as virtual replicas of real-world environments, enabling scientists to analyze conditions in extreme locations like Mars. The aerospace sector, much like NASA, is increasingly adopting AI and digital twins. According to Karen Willcox, director of the Oden Institute for Computational Engineering and Sciences, the stakes in this field are exceptionally high. Unlike many industries where AI deployment can be rapid and less risky, aerospace requires careful human oversight due to the potential consequences of errors. "AI is very powerful," she notes, "but it cannot replace human judgment." Digital twins are not a novel concept for NASA; the agency has utilized them since the Apollo missions of the 1960s. Today's digital twins, enhanced by AI, can predict issues, diagnose problems, and recommend corrective actions in real-time. This dual technology was critical when engineers at NASA's Jet Propulsion Laboratory developed Perseverance's travel route. The team fed data into an AI model, which generated a safe path by circumventing hazardous areas. The engineers then meticulously reviewed and refined this route before sending it to the rover. NASA's innovative use of digital twins extends to other projects, such as the James Webb Space Telescope. Due to its massive size, the telescope couldn't be fully tested in traditional thermal vacuum chambers. Instead, scientists created 3D video-based twins to monitor its intricate deployment process, which involves numerous potential failure points. Julie Van Campen, a systems engineer for the Webb project, highlights that the foundational work on these twins began in the early 2000s, well before the current capabilities of AI, but the integration of AI tools today is paving the way for the next generation of space exploration. Beyond NASA, the broader aerospace industry is keen on utilizing AI and digital twins for predictive maintenance, transforming traditional maintenance schedules into more efficient, data-driven processes. Companies like Airbus and Boeing are at the forefront, employing AI to enhance product development, manufacturing, and testing efficiencies. The Defense Advanced Research Projects Agency is also exploring AI’s role in speeding up testing for military systems while ensuring safety. However, as Willcox points out, the aerospace industry faces unique challenges that demand stringent verification and validation of AI systems. The risks involved necessitate a cautious approach, contrasting with sectors where rapid development is possible. As AI and digital twins evolve, they promise to enhance decision-making processes and real-time problem-solving capabilities, ultimately advancing the mission of exploring space without jeopardizing human safety or equipment.