On a barren stretch of California’s Colorado Desert this spring, a four-wheeled robot about the size of a coffee table drove itself roughly 16 miles (26 kilometers) while engineers walked behind it, mostly just watching. It crawled over sand and rubble in daylight and in the dark. The point was never the desert. The point was the Moon.
The machine is called ERNEST, short for Exploration Rover for Navigating Extreme Sloped Terrain, and NASA’s Jet Propulsion Laboratory detailed the results of its first field test on Thursday. The prototype is JPL’s answer to a problem that has quietly limited planetary exploration for two decades: NASA’s rovers are superb scientists and painfully slow drivers.
ERNEST covered that 16 miles over 37 hours of driving, spread across seven days, reaching speeds up to 0.6 mph (1 kph). It sounds glacial. In rover terms it is a sprint, an order of magnitude faster than the top speed Curiosity and Perseverance can manage on Mars.
“You could do a science road trip across the Moon — or Mars — with this vehicle,” said James Keane, a JPL planetary scientist working on lunar missions.
Why a faster rover matters
Since 1997, NASA has landed five rovers on Mars, and the two still working, Curiosity and Perseverance, move at a careful crawl. The caution is deliberate. Each is a one-of-a-kind machine worth billions, and a single bad rock can end a mission. But that same caution walls off the steep slopes, crater rims and shadowed ground where some of the most interesting science sits.
ERNEST is built to go there. At 4 feet (1.2 meters) long, it is far smaller than its SUV-sized predecessors, yet it can lift each of its mesh wheels on its own to step over obstacles that would stop Curiosity or Perseverance cold. An active suspension shifts weight between the wheels and switches gaits, letting the rover squirm, wheel-walk or climb, and its four steerable wheels let it drive in any direction, including sideways.
That is a clean break from the rocker-bogie suspension NASA has bolted onto every Mars rover since Sojourner. “We started by postulating that we could do better in designing a planetary surface robotic mobility system,” said Hari Nayar, the JPL principal technologist leading the ERNEST team. “While the rocker-bogie system has been very successful over the past 30 years, there’s been a lot of research in that time on mobility and understanding terrain interaction.”
Teaching a rover to think
Hardware was only half the job. ERNEST’s current design was finished in September 2024, but it still needed a human at the joystick. To cut that cord, the team turned to reinforcement learning, a form of artificial intelligence in which a robot improves by trial and error. JPL’s simulation lab built a high-fidelity virtual twin of the rover, fed it data from the real hardware, and ran thousands of hours of practice on a computing cluster, sometimes over a single weekend.
Then came the real thing: first an obstacle course of rubble piles and steep slopes in JPL’s Mars Yard, then the open desert in March. Engineers ran ERNEST in the dark on purpose. The long, knife-edged shadows of a desert night mimic the lighting near the Moon’s poles, where a future rover would go hunting for ice.
“This testing is helping us refine the mobility hardware and autonomy software to navigate extreme distances across a wide range of terrain and lighting conditions anticipated on the Moon,” said Issa Nesnas, a JPL principal technologist who leads autonomy work for a NASA concept for a long-range lunar rover.
ERNEST is one piece of a busy stretch for the agency, which is also mounting a first-of-its-kind robotic rescue of its aging Swift space telescope and pointing skywatchers toward a run of Moon-and-planet conjunctions this year.
None of this guarantees ERNEST ever leaves Earth. It is a testbed, and the team’s goal is to prove that a rover roughly twice its size could one day make a long lunar trek. The project began in 2022 on JPL’s internal research budget and now runs on NASA’s Mars Exploration Program. For now, the rover’s road trips end where the desert does. The ambition behind them does not.
