A legged robot that can size up several rocks in one go, instead of waiting for humans to guide every step, has passed an early test for future Moon and Mars missions.
Dr Gabriela Ligeza, a former PhD student at the University of Basel who is now a postdoctoral researcher at the European Space Agency, recently tested a semi-autonomous exploration system with colleagues. The team equipped a quadrupedal robot with measurement tools designed to investigate multiple targets quickly and collect mineralogical data.
The results, published in Frontiers in Space Technologies, showed the system could quickly investigate several targets, identify promising rocks, and return data useful for astrobiology and in-situ resource utilization, or “living off the land”.
Ligeza said the work tested a different model from current planetary surface missions, which tend to operate cautiously. On Mars, communication delays between Earth and rovers are typically between four and 22 minutes. Limits on uplink and downlink data transfer also mean scientists must plan operations in advance.
Rovers are also built to save energy, stay safe, and move slowly across hazardous terrain. That limits exploration to a small part of the landing site. According to the article, rovers typically travel up to a few hundred meters per day, making it harder to collect geologically diverse data.
The team instead tested a semi-autonomous robotic explorer that could investigate multiple targets one by one and collect data without constant human intervention. Their results showed that semi-autonomous robots fitted with compact instruments could speed up resource prospecting and the search for “biosignatures”, meaning evidence of life, on planetary surfaces.
Rather than stopping to investigate a single rock under continuous supervision, the robot could move to multiple targets and perform measurements at each location on its own.
The researchers asked a simple question: could a robot with a relatively simple scientific payload quickly study several targets and still produce meaningful scientific results? Their answer was yes. The study found that compact instruments could still meet the full scientific objective of identifying rocks relevant for astrobiology and resource exploration.
To test the concept, the team used the quadrupedal robot ANYmal. It carried a robotic arm fitted with two instruments: the microscopic imager MICRO and a portable Raman spectrometer developed for the ESA-ESRIC Space Resources Challenge.
The work involved the Robotic Systems Lab at ETH Zurich, ETH Zurich | Space, the University of Zurich, and the University of Bern.
The experiments took place in the “Marslabor” facility at the University of Basel. The site simulates planetary surface conditions using analogue rocks, regolith materials, and analog lighting conditions. In the tests, the robot autonomously approached selected targets, positioned its instruments with the robotic arm, and sent back images and spectra for analysis.
The system identified a range of rock types relevant to planetary exploration, including gypsum, carbonates, basalts, dunite, and anorthosite.
The article said many of those rocks are scientifically significant. It pointed to lunar-analog rocks such as dunite, which is rich in olivine and oxides, and anorthosite, which contains anorthite, along with oxides such as rutile, as possible signs of valuable resources for future space missions.
The team then compared two ways of working. One was a traditional single-target approach closely guided by scientists. The other was a semi-autonomous multi-target strategy in which the robot carried out measurements at several locations in sequence.
The semi-autonomous missions were faster. Multi-target missions took between 12 and 23 minutes. A human-guided mission took 41 minutes to complete comparable analyses.
The faster pace did not stop the robot from producing strong results. In one test run, all selected targets were correctly identified.
The study said this approach could let future missions survey large areas of planetary surfaces more quickly. Scientists could then review the incoming data and choose the most promising locations for closer study.
It also argued that future missions may not need to rely only on large, complex instrument suites. Agile robots using relatively simple instruments could rapidly scan the environment and flag promising targets for detailed investigation.
As space agencies prepare missions to the Moon, Mars and beyond, the study said semi-autonomous systems could help scientists survey larger areas in less time, while supporting resource prospecting and the search for possible signs of past life.
