University Professor uses robots in medicine, agriculture

Ph.D. candidate Dimitris Zermas plays soccer with Darwin the Humanoid Research Robot in Walter Library on Friday, April 3rd.

Alex Tuthill-Preus

Ph.D. candidate Dimitris Zermas plays soccer with Darwin the Humanoid Research Robot in Walter Library on Friday, April 3rd.

Brian Edwards

A University of Minnesota professor is using robotics to detect early signs of autism in children and reduce the amount of fertilizer running into rivers near farms. Department of Computer Science and Engineering professor Nikolaos Papanikolopoulos is developing robot and sensor technology to detect visual cues that could determine if a child has autism or a farmer needs to fertilize a certain area of crops. Papanikolopoulos is a part of the organization Robots and Sensors for the Human Well-bBeing , a group of faculty members and students from five universities across the country that develop and experiment with robot and sensor technology. âÄú[Papanikolopoulos is] internationally known, but you sit down with him and he describes his work in a way that anyone can understand,âÄù said Mike Bazakos , who is the managing director for the organization. With the group, Papanikolopoulos is working to develop unique ways of locomotion for robots âÄìâÄì like climbing ladders and staircases âÄî and then heâÄôs creating an automated system for running them. Papanikolopoulos presented his work at the College of Continuing Education last week, where he discussed the future of automation. He said unlike the self-driving Google car and AmazonâÄôs future drone delivery service, called Amazon Prime Air, his current work with robots focuses on assistance, instead of replacement. In 2009 , the UniversityâÄôs Medical School asked him to develop a technology to monitor early behavior signs of certain diseases. He said he was able to adapt his previous work of monitoring visual behaviors with the U.S. Department of Homeland Security in 2004 to fit the schoolâÄôs needs. Josh Fasching , a doctoral student in computer science, has worked with Papanikolopoulos for five years and is currently working with him on early autism diagnosis. âÄúWe are basically using computer vision tools to detect behaviors, like unprovoked hand flapping or head nodding,âÄù he said. During his presentation, Papanikolopoulos said automation canâÄôt replace the knowledge of doctors, but instead it can help distill large amounts of data and pinpoint areas that would then require further investigation. PapanikolopoulosâÄôs technology has applications in the agriculture industry, as well. Each October, Minnesota farmers pay millions of dollars for fertilizer and pesticide, which they then dump onto their fields, Papanikolopoulos said, and a lot of the chemicals end up in rivers. Bazakos said the research group is testing PapanikolopoulosâÄô technology with the Minnesota Corn Growers Association to learn how to more efficiently spread the chemicals more efficiently. âÄúThey have pre-determined times where an aircraft goes and sprays the whole farm, then three weeks later, sprays the whole farm again,âÄù Bazakos said. But, thatâÄôs often not the most efficient way to fertilize a field of crops, Bazakos said. Papanikolopoulos said the group also developed a solar- powered drone that can fly over farms and use sensors to detect yellow-colored corn leaves. The drone then disperses fertilizer only to those plants that need it. âÄúWhat we are trying to do is like precision surgery,âÄù Bazakos said. He said they are working on a ground robot to coordinate fertilization with the drone. Adam Czech, public relations manager for Minnesota Corn Growers Association, said farmers are always looking for ways to save money and maintain water quality. Still, farmers have yet to fully incorporate PapanikolopoulosâÄô technology, Czech said. âÄúThere is definitely a lot of interest, but right now we are in that really early stage,âÄù Czech said.