Cleaning up old nuclear waste sites around the country is a long, expensive and dangerous process - and autonomous robot research at the University of Nevada, Reno promises to help that process with a combination of advanced, intelligent, autonomous aerial and ground robots with a new level of perception, navigation and planning abilities.
The College of Engineering's Autonomous Robots Lab, under the direction of Assistant Professor Kostas Alexis, has completed a proof of concept for an aerial robot, a drone, that flies by itself in dark corridors looking for nuclear radiation and toxic chemicals.
"We've designed and built an aerial robot with multi-model mapping capabilities that includes inertial sensing, LiDAR, cameras with synchronized flashing LEDs, as well as sensors for radiation and chemical sensing to localize itself and comprehensively map its environment in very high quality," Alexis said. "Basically, it flies itself into a dark corridor, maps the area, including complex structures such as tanks or barrels, and simultaneously finds radioactive areas or toxic chemicals using a variety of sensors. It learns the environment - the space, the contents and the dangers - and reports back to us."
The intelligent aerial robot runs on algorithms the team programs to give it active perception, use what it needs to navigate, remembers important areas and allows production of high-resolution multi-modal 3-D maps of the area that show the level of radioactivity and toxic chemicals, if any. Their work is part of a National Robotics Initiative project funded by the Department of Energy to clean up the legacy sites of the Manhattan project that have been shuttered for decades.
The information will be used by the Department of Energy to build a clean-up plan based on what the autonomous robots locate, analyze and map.
"Specifically, we are hoping that we can enable the autonomous multi-modal mapping of the PUREX tunnels where multiple train cars are holding nuclear waste," he said.
The majority of the clean-up work is conducted by human workers in protective clothing with consequences of cost, inefficiency, exposures and inability to access many places that matter. There is a need for robotically and autonomously acquiring, integrating and utilizing radiological, chemical, thermal, spatial and visual data of the inaccessible facilities.
For the three-year program they have put together an interdisciplinary team of experts in perception, motion-planning, ground robots, micro aerial vehicles, and nuclear robotics. He is working with colleagues at the Robotics Institute of Carnegie-Mellon University.
"We believe that exciting new developments will arise in the intersection of these areas and themes," Alexis, a faculty member in the Department of Computer Science and Engineering, said. "We are one of several research groups participating in the DOE environmental management program to clean up 107 sites around the country."
Driver assisted technology for Reno buses
Alexis and his team in the Autonomous Robots Lab use their expertise in a number of cutting-edge projects, including locally in a project to develop driver-assisted technologies for use on buses for the Reno-Sparks area.
They are working through the University's Nevada Center for Applied Research in a collaboration with the Regional Transportation Commission, the Governor's Office of Economic Development, the Nevada Department of Motor Vehicles, the City of Sparks, bus company Proterra and the German research institute Fraunhofer. The project, Intelligent Mobility, includes integration of several systems. Alexis's contribution lies in the development of the perception systems and detection systems, as well as in the autonomous navigation and collision avoidance.
"The systems will perceive the environment - identifying signs, signals, avoiding vehicles, pedestrians and other objects and provide data for robust and safe navigation," Alexis said. "It's just like we, as humans, drive, perceive the world and make decisions."
The team has already begun testing the equipment using their own vehicles as they drive around the region, and the developed perception and real-time mapping technologies will migrate to the bus routes as the project progresses in the next year.
"We drive around at various times of the day and night so the system can localize itself and we can make a map," he said. "Our lab develops the multi-modal localization and mapping solution, fusing camera, LiDAR, GPS and inertial sensor data, for the autonomous navigation of the vehicles within the urban environment. The first prototype of this multi-modal mapping unit has been developed and successfully tested in both day- and night-time navigation. In addition, our lab develops small electric vehicles for rapid prototyping of our autonomous driving technologies."
Alexis and his team use their newly built indoor Autonomous Robots Arena, with support from the Vice-President of Research and Innovation's office, in the University's Applied Research Facility high-bay lab, a 25-foot wide, 50-foot long and 40-foot high space, where they run test flights and ground operations for their research. The padded floor and walls ease the bumps and crashes as they fly around obstacles and navigate without the use of GPS, the standard system for commercial aerial robots such as quadcopters or fixed-wing aircraft.
Autonomous Boat to Monitor Lake Tahoe near-shore
The team heads outdoors for other research as well, such as an ongoing autonomous boat project in collaboration with NAASIC and the College of Science's Global Water Center to scan the bottom of Lake Tahoe for algae growth and other environmental impacts on the lake's nearshore habitat.
Alexis has run successful experiments surveying the lake bottom using high-definition cameras mounted on the underside of the autonomous boat. The interdisciplinary team of scientists and engineers is focused on developing a low-cost, autonomously operated surface water drone that is able to navigate and monitor the nearshore habitat, and use it as a model for developing similar programs around the world.
Coming to the University in 2015 after working in the world-renowned Autonomous Systems Lab of ETH Zurich in Switzerland, Alexis was excited to start his own lab to design, build and program aerial, marine and ground based autonomous robots.
"I truly enjoy the design and development of new robots and I have a personal preference on aerial systems with advanced dynamic properties," Alexis said. "The University offered me a good start-up package, to immediately begin building a lab, with resources for equipment and personnel. With NAASIC and other activities going on here, this is a good and very solid community for research and development."
Alexis specializes in autonomous robots with an emphasis on aerial robots, solar-powered UAV, augmented reality, marine and ground robots, robotic perception systems and the algorithms to enable intelligent behaviors. His goal is to make important contributions into making robots fully autonomous.
"The aerial robotics community has moved from flying autonomously with GPS to flying without GPS and navigating with in a safe collision-free manner," Alexis said. "My next step is to establish cognizant robotics, where we can do complex missions autonomously without prior knowledge, without people commanding and running the mission."
Those complex missions employ sensors and are scalable to do many types of missions large and small, such as what Alexis and his team are undertaking now. His team continues to grow, with his senior researcher Christos Papachristos, doctoral students Shehryar Khattak, Tung A. Dang, and Frank Mascarich; Graduate Research Assistants Tyler Sorey and Niki Silveria as well as several undergraduate researchers.
"Students are an important part of this lab," Alexis said. "there are about 120 students at the University studying robotics. I'm holding a robotics day, like an open house, this spring semester (March 31, 2017) to introduce the lab to students, faculty and the community, hoping to attract the best talent to do research in our lab."
"Kostas's Autonomous Robots Lab is an impressive facility with state-of-the-art technology, an impressive research group and major projects of national/international importance," Manos Maragakis, dean of the College of Engineering, said. "This lab exemplifies progress and excellence in an area that we strategically decided to focus on five years ago. This is also an area of great interest to industry and entrepreneurs regionally and nationally."
Alexis obtained his doctorate in the field of aerial robotics control and collaboration from the University of Patras, Greece in 2011. His doctoral research was supported by the Greek National-European Commission Excellence scholarship. Subsequently, he moved to Zurich, Switzerland where he worked as a Senior Postdoc Researcher at the Autonomous Systems Lab of ETH Zurich.