We’ve come a long way from consumer robots like the Roomba or Furby. Beyond entertainment or completion of simple tasks, we’re now seeing a wave of new technologies that combine artificial intelligence and robotics to optimize intricate processes throughout a diverse set of industries.
In the world of manufacturing, robots are a crucial piece of the future, helping to automate processes and reducing how much time or money is required to produce quality results.
At the 2019 AeroDef Manufacturing Conference, produced by SME, current and former USC Viterbi graduate students from the Department of Aerospace and Mechanical Engineering (AME) presented innovative research concerning how best to harness robots to improve intricate manufacturing processes.
AeroDef Manufacturing annually gathers industry professionals, researchers, and academics to foster innovation toward improvement of manufacturing practices. From the pool of submissions, only 24 teams were invited to present at the SME AeroDef Manufacturing Poster Challenge in Long Beach, CA. AME students took home the first and third place prizes in the graduate-level contest.
The first place award, along with a $750 prize, was given to the project titled, “Smart Robotic Assistants for Surface Finishing,” completed by AME Ph.D. students Ariyan Kabir and Rishi Malhan, AME graduate Aniruddha Shembekar—currently a research engineer at the Center for Advanced Manufacturing (CAM)—and Brual Shah, who is a research scientist at CAM. The team was advised by Smith International Professor in Mechanical Engineering and Computer Science, Dr. Satyandra K. Gupta, who is also the director of CAM.
The first-place team demonstrated its collaborative surface finishing system, devised so human operators can work on high-level decision-making, while robot assistants complete labor-intensive, low-level finishing tasks. Robots are programmed based on user inputs and task requirements, and by planning algorithms, which automatically compute paths for the robots by using the computer-aided design (CAD) model of the part in question.
“Our task and motion planning algorithms enable robots to generate their own instructions. Our self-directed learning algorithm determines the near-optimal process and robot trajectory parameters by conducting a small number of experiments. These artificial intelligence(AI) modules significantly reduce the robot programming time and improve the efficiency of the finishing system,” says Kabir.
Third place and $250 was awarded to the project titled, “Realizing Supportless Extrusion-Based Additive Manufacturing through Use of Robots,” highlighting the work of AME Ph.D students Prahar Bhatt and Rishi Malhan. The team, also advised by Dr. Gupta, looked at how to use robots to eliminate the need for support structures—which increase material waste and build time—by using gravity to our benefit instead of opposing it using support.
“Though the overall system is computationally complicated, once the algorithm is developed it can be used by anyone to replicate it. This technology has a future to be adopted in commercial applications especially in metal AM where the removal of support structures is difficult,” said Bhatt about the project.
The AeroDef Manufacturing Conference was held from April 29 to May 2. Competing teams presented their posters on April 30 and winners were announced on May 1.