Drilling on Autopilot

September 23, 2004 / Posted by: Yael Kovo

Carol Stoker is the principal investigator for the Mars Analog Research and Technology Experiment (MARTE). MARTE has just begun its second field season drilling into the subsurface near the headwaters of the Río Tinto in Spain, searching for novel forms of microbial life. In a four-part interview with Astrobiology Magazine Managing Editor Henry Bortman, conducted just before Stoker left for Spain, she explained what MARTE hopes to accomplish. In this fourth and final part, Stoker talks about MARTE’s technology objective: developing a fully automated drilling and life-detection system.

Astrobiology Magazine (AM): In your first field season last year, and again this year, your focus is on conducting science-based operations, searching for subsurface life near the Río Tinto in Spain. Next year, though, in your third and final field season, your focus will be shifting to the technology side of your project. You’ve set yourself a difficult task for year three: conducting a drilling operation that doesn’t require human intervention.

Carol Stoker (CS): Yeah. Our project has been developing a first-ever completely robotic system to do a coring drill and sample analysis autonomously. We took as guidance the question: How do geologists and biologists figure out how to find life in the subsurface and can we automate that? So we are building a robotic coring drill that will bring cores out of the ground and break them off and put them into a core-sample-handling system. The core-sample-handling system will then expose the cores to a suite of remote-sensing instruments that will diagnose their mineralogy in the same way that the instruments on the Mars Exploration Rovers diagnose the mineralogy of the surface samples there. Then a science team will look at that data and decide where would be a good place to do a subsample to look for life. Then an automated system will cut off chunks of that core, grind it up in a rock grinder and put it into a suite of life-detection instruments.

That whole process will be done autonomously. So there’s a lot of stuff we’re building that’s never been built before. That will be demonstrated next year in a simulation of a Mars drilling mission at Río Tinto, at the same field site where we’re doing our science activities.

AM: How autonomous do you expect next year’s operation to be?

CS: We’re building four new robotic systems that have never been built before – and that’s just the mechanical systems. Then you have all the software to make it all work. My experience with rover operations tells you that you can easily spend 4 times the amount of money you spend on the hardware – 10 times the amount of money you spend on the hardware – getting the software to run flawlessly. That is the roughest job for any autonomous system.

Our project was funded at the level that we could pretty much do the hardware but we didn’t really have the funding to do the software, at least not at the level that it would require to have a system that was bombproof, autonomous, that you could just stick out in the field and control it completely from somewhere else. So our intention is to have a team at the borehole that will essentially supervise the autonomy. The system will run autonomously, but the team at the borehole will be there in case anything doesn’t work.

Commands will be sent form a remote operations center in Madrid. But the intention is to demonstrate the ability to do a robotic operation of the drill. To get the autonomy to be perfect, to get all the software debugged, is many man-years of work. There’s nothing in principle that would be new and unique about that, whereas the robotic systems, the mechanical systems that we’re building are new in principle; that stuff has never been designed before.