The moon contains untapped resources that humans may eventually mine and exploit. Agencies such as the European Space Agency (ESA) are preparing to delve into our celestial neighbor to identify these minerals. To effectively explore the lunar surface, a team of Swiss scientists at ETH Zurich envisioned sending not just one rover but a coordinated fleet of various vehicles and flying devices that could work together.
The power of a team is greater than the sum of its parts - a three-legged robot is tested in a quarry in Switzerland. Image source: ETH Zurich/TakahiroMiki
The researchers equipped three ANYmal, a legged robot developed by ETH, with a range of measurement and analysis instruments that could potentially make them suitable detection devices in the future. They tested the robots on various terrains at the European Space Resources Innovation Center (ESRIC) in Switzerland and Luxembourg. A few months ago, a Swiss team won the European lunar robotic competition together with their German colleagues.
The competition involves finding and identifying minerals on a test site modeled after the lunar surface. In a recent paper published in the journal Science Robotics, scientists describe how they used teams of robots to explore unknown terrain.
"There are two advantages to using multiple robots," explains Philip Arm, a doctoral student in the research group led by ETH Professor Marco Hutter. "A single robot can perform specialized tasks simultaneously. Furthermore, thanks to redundancy, a team of robots is able to compensate for the failure of its teammates. In this case, redundancy means installing important measurement equipment on multiple robots. In other words, redundancy and specialization are opposing goals. To obtain the advantages of both, one must find the right balance."
Swiss engineers are building robots suitable for future moon missions to search for minerals and raw materials. To ensure that the machines can continue to work even if one of them fails, researchers are teaching them to work as a team. Source: University of Zurich/Central Information Technology Center - MELS
To solve this problem, researchers from ETH Zurich, the University of Basel, the University of Bern and the University of Zurich equipped two of the legged robots with experts. One of the robots was programmed to be particularly good at mapping terrain and classifying geology. It uses a laser scanner and several cameras, some of which can also perform spectral analysis, to gather initial clues about the mineral composition of the rocks. Another expert robot learned to accurately identify rocks using a Raman spectrometer and microscope camera.
The third robot is a general-purpose robot: it can both map terrain and identify rocks, meaning it has a wider range of tasks than a specialist robot. However, its equipment means it is less precise at performing these tasks. "This makes it possible to complete the mission if either robot fails," Arm said.
The jury was particularly impressed by researchers at the ESRIC and ESA Space Resources Challenge competition who built redundancy into their exploration systems to make them resilient against potential failures. As a reward, the Swiss scientists and their colleagues from the FZI Information Technology Research Center in Karlsruhe received a one-year research contract to further develop the technology. In addition to legged robots, this work will also build on FZI researchers' experience with such robots to develop wheeled robots.
Hendrik Kolvenbach, a senior scientist in Professor Huth's research group, explains: "Footed robots like our ANYmal excel in rocky and steep terrain, for example when climbing down volcanic craters. Robots with wheels are at a disadvantage in this situation, but they can move faster on less difficult terrain. Therefore, in future missions it makes sense to combine robots with different modes of locomotion. Flying robots can also join the team."
The researchers also plan to increase the robot's autonomy. Currently, all data from the robots flows into a control center, where operators assign tasks to individual robots. In the future, semi-autonomous robots could assign certain tasks directly to each other, with operators able to control and intervene.