New rover system can make its own decisions and… improvise
Rovers are great pieces of engineering meant to function in even the harshest environmental conditions. But whilst the most recent versions of these scientific labs on wheels can perform some operations autonomously, there are still many situations they can’t handle on their own. Say a rover needs to travel a very long distance on the surface of the Moon or Mars. On its way, it might encounter unforeseen environmental hazards or even discover interesting features that arouse its scientific curiosity. There are currently two possible ways to deal with such situations: teleoperation or on-board sequencer plans improvised by an army of mission operators and scientists back on Earth. But the ADE (Autonomous decision making in very long traverses) consortium proposes an alternative: a demonstrator relying solely on autonomous decision-making by the rover itself. “We proposed and successfully proved a system with an on-board planner that can decompose high-level goals – such as moving a sample from position A to B or taking an image at a given position – into actions fully independent of human control,” says Mariella Graziano, executive director of Flight Systems and Robotics at GMV. “We achieved this thanks to an autonomous navigation capability, a scientific agent to search for patterns of interest and analyse them, a novel fault detection, isolation and recovery (FDIR) system, a soil traversability analysis, and many other challenging features.” The new rover is an impressive and complex system of systems, but what truly stands out – and is even revolutionary in project coordinator Jorge Ocon’s own words – is its mission planning system. “The approach that consists in adding a dynamic planner on board and combining it with the other rover subsystems is a clear breakthrough. It’s a whole new way of commanding and managing space missions that will increase mission performance, reliability and optimisation of on-board resources in unknown and critical environments,” he notes. To put it simply, ADE is pushing a paradigm shift in which ground station operators tell the robot what to do, but let it decide how to do it. This is all expected to reduce the time required to perform rover operations on Mars or the Moon, increase scientific return and improve rover resilience.
Destination The Final Frontier… but not only there
With its autonomous decision-making, the ADE system could prove key to the success of future human exploration of Mars and the Moon. The astronauts who will eventually set foot there will need all the intelligence they can get to guarantee their survival, and that implies precise knowledge of the planetary environment, location of resources and even places where settlements can be established. But it can have its uses on Earth, too. “The capabilities of ADE can be used in nuclear or mining environments, during search and rescue missions or in underwater operations, and proving this potential was a secondary objective for the project,” Graziano adds. GMV successfully developed a demonstrator which can characterise an area autonomously, detect and delimit areas of high radioactivity within a nuclear power plant, and spot leaks that could require human intervention. With the project now completed, GMV will continue to mature its new technology. They hope to see it used both in future space exploration missions and during sensitive missions here on Earth.
Keywords
ADE, Moon, Mars, space exploration, rover, autonomous, mining, nuclear