Researchers have demonstrated that a legged robot can autonomously analyze multiple rocks in a single mission, completing surveys far faster than traditional step-by-step control.
The result shows that planetary exploration can move beyond stop-and-wait workflows, allowing machines to scan terrain and return useful data without constant human direction.
Autonomous robots on Mars
Inside a controlled Mars-like test environment, the robot moved across simulated dust and rock fields, stopping at several targets within a single run.
Working in that setting, Gabriela Ligeza at the University of Basel showed that the system could approach each target, deploy its instruments, and return meaningful scientific readings.
Across repeated trials, the robot consistently identified key rock types while operating without continuous input between measurements.
That performance establishes a faster but less supervised mode of exploration, raising the need to balance speed against precision in how future missions are planned.
Communication lag time
Earth-to-Mars messages can take up to 22 minutes one way, so communication delays turn each new rock into another planning cycle.
Because teams must review data and uplink commands, traditional surface work usually advances one target at a time.
Instead, the new trials let scientists pick several rocks first, then let the robot walk, measure, and report back.
That sequence attacks rover science’s slowest part, the human pause between one interesting stone and the next.
Cutting-edge diagnostic sensors
At the arm tip sat a close-up imager, while a Raman spectrometer, a laser tool that fingerprints minerals, rode above.
Close views revealed cracks, grains, and textures, while Raman readings flagged mineral chemistry that appearance alone could miss.
Together the tools identified a range of rocks tied to water, volcanic activity, and potential resources across Mars and Moon stand-ins.
That pairing mattered most when one tool stumbled, because blurred pictures or dusty surfaces did not always fool both instruments.
The robot delivers speed
Across four semi-autonomous Mars-style runs, the robot finished surveys in 12 to 23 minutes, while a guided run took 41 minutes.
During the best fast run, the robot identified all three chosen targets, matching the slower approach on basic success.
Meanwhile, the guided mission gathered more useful data because humans could inspect each result and request extra measurements.
That trade-off defined the core result: autonomy bought speed, while constant supervision still bought precision and higher total yield.
Supervision can boost detail
When scientists stayed in the loop after every stop, they could judge image quality and request another reading immediately.
In the lunar test, that freedom produced ideal close-up images for all three targets and six Raman measurements.
One reading came back weak, showing how mixed mineral layers can confuse compact instruments even when aiming succeeds.
Human review mattered less for walking than for deciding when a target deserved another try before moving on.
Robots on Mars need legs
Wheels excel on smoother routes, but many prized lunar and Martian deposits sit on steep ground where footing matters more.
ANYmal, a four-legged robotic platform designed for rough terrain, uses onboard sensing to map the ground and choose stable footholds while it walks.
That matters for future prospecting because ice, metals, or rocks that preserve life traces often lie near cliffs or slopes.
A faster walker that still carries science tools could scout risky places before larger missions commit power and time.
Targets for future missions
Several materials in the test were stand-ins for rocks that mission planners already care deeply about.
On Mars-like ground, gypsum and carbonates matter because water helped form them, making them good targets for biosignatures, clues to past life.
Across lunar stand-ins, certain rocks pointed to materials that could provide oxygen or useful metals for future missions.
Those examples show why a quick first pass matters, since the hard part is not finding a rock but choosing well.
The compact gamble
Compact instruments usually trade depth for speed, so the real test was whether a smaller payload could still sort rocks well.
That problem was laid out in a Frontiers editorial that explained the team’s goal in plain language.
“Our research question was whether a robot equipped with a simple scientific payload could quickly study several targets while still delivering meaningful scientific results,” said Ligeza.
The result wasn’t flawless accuracy at every step, but it delivered enough consistent sorting to make rapid scouting a realistic first pass.
Autonomous robots and Mars exploration
The next upgrade is not a bigger robot but smarter checking, so it can catch blur and retake weak shots.
Future versions could flag targets by color, texture, or shape, then send back a short list when bandwidth is tight.
NASA already lets Perseverance use onboard artificial intelligence to choose targets and drive efficiently, but not to plan a multi-target campaign.
One scout that can walk, inspect, and move on begins to break the stop-and-wait rhythm shaping planetary fieldwork.
The result does not replace slower, expert-guided science, but it offers a workable way to cover more ground.
The study is published in the journal Frontiers in Space Technologies.
Image Credit: Dr Tomaso Bontognali
—–
Like what you read? Subscribe to our newsletter for engaging articles, exclusive content, and the latest updates.
Check us out on EarthSnap, a free app brought to you by Eric Ralls and Earth.com.
—–