← Back to NASA Technology Projects
Completed TRL 3 (started at 3, targeting 4)
Non-geometric terrain hazards are mission-enders for autonomous ground vehicles on Earth and on other worlds. Spirit was mired in soft soil on Mars and Lunokhod became stuck in regolith on the Moon. Autonomous ground vehicles must detect and react to off-nominal conditions to limit risk and maximize performance. This research investigates multiple sensing modes for short- and long-range terrain property measurement by incorporating long-range sensing and receding horizon control to ensure safe operation during autonomous planetary traverse, allowing robots to perceive and react to non-geometric terrain hazards, such as soft soil, before engaging them.
This research takes a two-pronged approach, incorporating long-range sensing and receding horizon control to ensure safe operation during autonomous planetary traverse. It will investigate and develop detection of non-geometric hazards with multiple sensing modes to ascertain terrain properties at range. Self-supervised learning maps multi-mode long-range sensor data to terrain properties, measured by direct-contact instruments, and to observed vehicle performance. Receding horizon contingency planning applies learned models of vehicle-terrain interaction to maintain stability and safety.
This research will enhance knowledge and understanding of robotic driving in unstructured terrain.
This activity will investigate multiple sensing modes that could provide non-geometric terrain hazard estimation from a distance. This will enhance knowledge and understanding of robotic driving in unstructured terrain. Good methods exist for sensing geometric obstacles; the limitation now is non-geometric hazards. Human drivers can infer non-geometric terrain properties well from prior experience, to pilot vehicles near their physical limits. State-of-art computer systems can infer some non-geometric properties at short range but are limited. Research impacts the fields of artificial intelligence, robot perception, and planetary robotics
Automated perception and reaction to non-geometric hazards have defied autonomous robots. Despite great advances in mechanism, electronics, and software reliability, robots cannot achieve systemic reliability until they perceive and react to terrain hazards before engaging.
When coupled with speed and dynamic effects, hazard assessment is a major barrier to infusion of automated safety features in passenger cars. Fast, robust, predictive assessment of hazard will enable driving, working and exploring near the limits of machine capability.
The fully matured technology could provide future planetary rovers, such as Mars 2020, with the capability to sense non-geometric terrain hazards from a distance, hazards that are not detectible based purely on visually inspection, allowing them to avoid being trapped in loose soil.
Terrestrial search and rescue robots could potentially benefit from this technology as well.
Listed on TechPort itself — the most direct way to ask about this specific project.
This is early/mid-stage (TRL 3) — the most realistic path in is NASA SBIR/STTR, which funds small businesses and research institutions to develop technology aligned with NASA's needs (equity-free, phased funding). Check whether a current SBIR/STTR solicitation topic overlaps with this project's technology area, or contact the project directly (above) to ask.
None of these are guaranteed paths for this specific project — TechPort itself doesn't have an "apply" button. Reaching out to the contact(s) above with a specific question is usually the fastest way to find out what's actually open.