Autonomous robots rehearse for Mars and the deep sea

Freezing cold or hellish heat; extreme pressure or vacuum; toxic atmosphere; deadly radiation; eternal darkness: The deep sea and outer space are not exactly inviting. And since humans aren't necessarily made for such extremes, the idea of the project is that robot teams will go instead. The teams consist of individual specialists that are networked together and support each other. The demo mission was intended to show what this might look like on a space mission, so 55 scientists and engineers–among them 12 from the European Space Agency ESA–met in Sicily on Mount Etna..

Explore, control, build a base

"Etna is an active volcano. That means the rock in this environment has not yet been heavily altered by environmental influences. It resembles the rock on the moon and Mars," says Armin Wedler, explaining the choice of mission location. "Technologically, this creates some challenges, for example, for the locomotion of the robots. And scientifically, it offers the opportunity to practice here on Earth with analysis procedures for later missions to other celestial bodies.

Even robot teams need a plan

But how do you get a wide variety of robots to work together and, if necessary, not be thrown off course by the interjections of a human operator? Sören Hohmann at the KIT Institute for Control Systems has been researching this successfully for many years. Esther Bischoff from his team explains, "If you have different robots that are to perform different tasks within a mission, that has to be planned." That sounds simple at first. But in most cases, certain boundary conditions have to be met for this to happen. For example, there are tasks that may only be started if another one has been completed first. "And if the preferences of the human project participants are also to be taken into account, this planning is better handled by a computer algorithm." The young scientist has designed one such algorithm for the ARCHES project and demonstrated its capabilities on Etna.

"We used a huge touch display that showed a terrain map along with the current position of the different robots," she says. "For this, we designed a planning and coordination algorithm and developed the interactive user interface used to interact with humans." In the simulation, it was then possible to define all the tasks that the robots would later perform. The operator could also specify whether a very specific robot was intended for a task or whether the system should decide freely. And during the mission, the operator was free to intervene at any time. For example, new tasks could be added or certain robots could be assigned a different task. "We brought everything with us to Etna, set up the simulation there and presented it to colleagues from the other centers," explains the electrical and information engineer. "The different tasks could be defined in the user interface and were then processed in the simulation."

Extreme worlds challenge science

Among the colleagues listening intently to the presentation was Sascha Flögel. The geologist and paleontologist conducts research at GEOMAR. "I was there as a spectator this time," he recounts. "Together with AWI, we from GEOMAR already had our baptism of fire two years ago." referring to the deep-sea demo mission that took place in October 2020 in the Eckernförde Bay. Due to the pandemic, only a small team of marine researchers was involved in setting up an autonomous underwater measuring network and successfully demonstrating the cooperation of seven mobile and stationary measuring platforms.

"It's quite different when the mission control center is a fixed building with a connection to the supply and not a swaying ship all alone in the ocean," says Sascha Flögel, looking at the control room in the hotel on Mount Etna. "And then also to see the systems and their interaction up close was a very exciting experience."

After all, even though both are extreme environments, space and deep sea sometimes present researchers with very different challenges. For example, in communication. "Even though there are runtime delays in transmission, in spaceflight the mission control is usually in constant contact with the robotic system," explains Sascha Flögel. "In ocean research, it's different. Once the systems are set down in the ocean and not connected to the ship via a cable or acoustically, then you don't have access to them for the duration of the mission." That means researchers usually don't find out whether everything went well until 12 or 14 months later. Whether all the measurement systems have worked and whether the data are all there.