The Problem: There is no GPS (absolute positioning) on Mars – nor on any other extraterrestrial body of scientific interest. All existing navigation solutions for rover/robotic platforms provide only relative navigation capabilities, which drift off over time and are not sufficiently accurate for many scientific missions of interest. Future autonomous robotics missions will require an absolute positioning (GPS-like) capability, particularly for longer missions that last on the order of months or years (e.g., the Mars rover missions).
The Solution: OKSI proposes an Absolute Positioning System (APS) that would provide a day-night GPS-like navigation solution for ground-based robots. The solution is based upon matching real-time imagery captured by the rover to a pre-existing georeferenced satellite imagery database. Specifically, a rover platform would leverage available lofted on-board wide FOV cameras (e.g., Figure 1) to construct a full or partial birds-eye view NADIR projection of the surrounding scene. Then, the rover would take this projection of the surrounding scene and would match it to a georeferenced satellite database for the planet or body of interest. Once the real-time image is matched, it will be georeferenced via the database and then the pose of the platform can be extracted.
NASA’s next-generation planetary exploration missions will benefit from the technologies we are proposing – the OptoKnowledge APS directly addresses the Guidance, Navigation, and Control (GNC) need for robotic platforms such as terrestrial rovers. The APS would support the NASA HEOMD directorate by developing core transportation elements required for beyond LEO human exploration. The APS supports the Science Directorate by providing new sensors and instrument capabilities that provide low-cost and low-SWaP.
The proposed APS would provide a significant capability as a software add-on for absolute positioning to semi-autonomous vehicles such as Tesla’s fleet, as well as the several competitors in the autonomous ride-share market, such as Uber, Lyft, and Cruise2. These companies do not yet have true absolute navigation that enables lane-level autonomy.