We utilize cloud computing and high speed networks to make industrial robotics more flexible and adaptable. Using robotics, cloud computing, and 3D sensors, we are developing tools to help automate more complex aspects of manufacturing.
As computing power becomes more readily available in the market, we’re constantly looking for ways to utilize it. Our goal is to assign tasks to robots that are mixed-part and low-volume. This means that we give these robots tasks that require them to be aware of their environment when deciding the next appropriate move.
We have created two distinct applications with this concept in mind. In project Gilbreth, we use a UR10 robotic arm with 7 joints. When communicating with a 3D sensor, it is capable of distinguishing multiple different items (gears, pistons, etc.) on a moving conveyor belt. It is then able to position itself appropriately over the targeted piece, grab it, and move it to an appropriate location. Project Godel uses similar components, and is tasked with blending metals. This uses a robotic arm with 6 joints to blend and smooth metals, a task that continues to injure many workers to this day. We’ve complimented the ability to blend metal with the ability to ensure smoothness by using a Keyence Laser Scanner to detect any bumps or rough surfaces, which then leads to correcting the issue before blending has ceased.
Acknowledgments: We thank NSF for the UVA grant 1531065 and UTD grant 1531039. We thank Don Hicks, UTD, for introducing us to SwRI. Finally, we thank US-Ignite for offering us the opportunities to present our applications in demonstrations and/or posters at the US-Ignite Application Summit in 2016 and 2017.
Team Information: University of Virginia, Southwest Research Institute, The University of Texas at Dallas, Malathi Veeraraghavan