NASA rovers have revealed fascinating details and close-up images of Mars and the moon, but exploration of other planets is slow. Most rovers are remotely controlled on Earth, and their careful, energy-efficient design limits how much space they can cover.
Now, a new study from the University of Basel, ETH Zurich, and the European Space Agency, published Tuesday in the journal Frontiers in Space Technologies, explores how “self-thinking” robots can help us search the planet’s surface much faster.
Because the rovers are built and guided in certain ways, observations are often limited, making it difficult to collect data over a large area of the planet. It also takes time to transmit data over large distances between Earth and other celestial bodies.
“Rovers are designed to be energy efficient and safe, and to travel slowly in dangerous terrain,” the researchers said in a statement. “Here, we explored a different approach.”
They experimented with a “semi-autonomous robotic explorer” that could investigate multiple targets and collect data without being told where to go.
“The results showed that autonomous robots equipped with integrated instruments can greatly accelerate the observation of resources and the search for ‘biosignatures’ (ie, evidence of life) in planetary environments,” the researchers wrote.
So instead of humans controlling the robots by using a remote control from a long distance, the researchers found that a robot equipped with the right tools can navigate the surface and collect data from different locations on its own.
Self assessment
The researchers used a quadrupedal robot with a robotic arm. The robot, which they planned to operate without humans, carried a special camera and spectrometer, designed to identify various objects.
“The autonomy (of the robot) is limited to the execution of pre-defined instructions, including walking, waypoint navigation, instrument deployment and data retrieval,” the researchers wrote in the paper.
To test the robot, the team used the Marslabor facility at the University of Basel in Switzerland. The facility simulates the surface conditions of planets with rocky surfaces and special lightning.
The robot moved through the area to selected locations, such as certain types of rocks hidden in the area. It then used its robotic arm to send scientific instruments and return images and data for analysis.
The researchers compared the new strategy to a more traditional method: scientists directing the robot to inspect one target at a time.
The researchers found that most targeted missions took an autonomous robot 12 to 23 minutes, while a human-guided mission took 41 minutes to perform the same assessment and analysis.
The direction of the moon
While most rovers use wheels to get around, using a legged robot could be useful, the researchers wrote, as some lunar resources are located in hard-to-reach areas with high crater walls, which could be a challenge for wheeled robots.
NASA is also working on autonomous rovers that can explore and navigate the Moon in a small manned manner. And while the current Artemis II mission will not use a lunar rover, these robots could help pave the way for continued human presence on the moon.
In December, NASA used AI to successfully program the Mars rover Perseverance course. The team used Claude AI’s Anthropic models in two demonstrations. The rover navigated obstacles on the Martian surface and traveled about 1,500 feet without ejecting.
