Convergent Manufacturing using Multiple Industrial Robots

Funding Organization: 

Advanced Robotics in Manufacturing (ARM) Institute

Principal Investigator: 

John T. Wen

Co-PIs: 

Johnson Samuel, Agung Julius, Santiago Paternain, Glenn Saunders

Dates: 
November 1, 2022 to October 31, 2023
Amount: 
$575,000
Grant Status: 
Current
Abstract: 

To program an industrial robot to follow a complex path, the most common approach is to use it like a machine tool and focus only on the robot tool center point (TCP) motion. The desired tool path is decomposed into straight line tool motion segments and then blended to avoid sharp turns. Currently the way points, blending zones, and trajectory speed are manually tuned to ensure high speed and high accuracy tracking of such paths. The motivating use case and demonstration facility is provided by GE Global Research based on their cold spray coating application, a deposition process where small metal particles are forced through a nozzle at high speeds before impacting a part. The problem posed by this application is: How to autonomously create a high-speed and high-precision curvilinear robot tool trajectory on a complex curved geometry using industrial robots with redundant degrees of freedom. The objective of this project is to solve this problem by developing an optimization approach to decompose a given path to a sequence of robot motion primitives, moveL, moveC, moveJ, with specified parameters (blending zones, path speed, robot pose and redundancy resolution), to achieve minimum cycle time with guaranteed tracking accuracy. The optimization algorithm will be evaluated and demonstrated in simulation and on physical testbeds. Simulation-based data will be used to train machine learning tools to reduce the optimization computation time. The software implementation will be open source and available in ROS and Robot Raconteur.