Robots to Provide Autonomous Soil Sensing

In environmental engineering, there's huge value to infield real-time sample analysis. Watch this video to learn about how a team at Carnegie Mellon University is building robots to provide this information.

“We are now looking to move beyond this proof-of-concept stage into more concrete objectives and demonstrations,” said Greg Lowry  , professor of civil and environmental engineering.



Transcript

00:00:13 Greg: So right now we're working on robots for  environmental sampling. In particular, we're   working on robots that can characterize large  impacted sites very inexpensively and rapidly.   So it doesn't just have to be soils, it could also  be anywhere you don't want to send people, you can   send a robot to sample and collect information  for you and keep people out of harm's way. Aaron: Another advantage is we can get more data  and higher-quality data that's less prone to human   error, and we can send the robot out for hours  collecting sample after sample after sample. Greg: In environmental engineering, there's huge  value to infield real-time sample analysis because   right now it takes forever to collect a sample,  send it to lab, get results back, make decisions.   We're trying to build robots that can provide that  information in real time so that the robot can   decide where to go next and to do the job better,  sample things faster and more inexpensively.

00:01:09 Vivek: What makes it different is we have a  portable x-ray force and sensor. You can scan   a sample and figure out what kind of elements are  there. The robot is designed such that it can go   in challenging environments, and it also has an  arm that helps lower the sensor to the ground   because the sensor has to be almost touching  the sample for it to make a valid measurement. Hairong: So this sensor have several  advantages. First, it's, like,   the sensing time is very short. It can get the  result in several minutes. And then the other   thing is, like, you don't need a complicated  sample preparation for this kind of technique. Aaron: The impact of these studies will be  better characterization of these sites that   will allow us to remediate any contaminants  that are in the soil. We've started working   with both robotics companies, as well as  environmental engineering companies to try and  

00:02:00 transfer some of this new approaches into the  field and those companies have built their own   versions of these robots and deployed them at a  customer site, which is very exciting to see them   actually deployed. One of the unique things about  this collaboration is how interdisciplinary it is. Greg: Neither Aaron nor I could  do this independently. It has to   be a collaborative effort in order  to make this work, and I think that   is a unique collaboration and it can  only be done in a collaborative mode. Aaron: And really it takes all of these different  disciplines coming together to build this kind   of complete system. On the robotics side, one  of the challenges is just getting through the   terrain. And different sites may be rocky  or have dense vegetation or sandy soils,   and so that affects the mobility  and how the robot is constructed.

00:02:50 Greg: This project has opened up a host of  new opportunities for us doing environmental   chemistry and sensing to apply it to various  scenarios and not just soil contamination,   but it could be tank inspections, it  could be detecting invasive species. Aaron: I think this research really shows how  robots can help us handle different environmental   challenges and do so in a way that reduces  the risk for humans involved in the process. Greg: I would say that robotics, automation,   and AI can improve environmental  engineering and remediation. Aaron: I think the key takeaway  here is that robotics is now at a   point where the systems can get over  this kind of challenging terrain and   really perform useful work and help keep  humans out of these challenging situations.