Daniel Gleeson, Electrical Engineering

​Half-Way PhD Semimar
Title: Industrial robot trajectory optimization

​Daniel Gleeson is a PhD student at the Division of Systems and Control
Examiner is Professor Bengt Lennartson
When implementing a highly automated production line, which is often the case in for example the automotive industry, there are typically a large number of industrial robots whose movements need to be determined. The goal is to perform a given set of tasks, such as spot welding or laying down sealing material. The first step is often to solve path planning and scheduling problems, to find collision free paths to perform all tasks and dividing and ordering the task execution in an efficient manner. The solutions returned by path planning algorithms are in the work presented here taken as a starting point, to be further refined, optimized or simulated. The goal is to seamlessly automate the full process, from a given set of tasks to be performed to robot code that directly can be used in a factory setup.

The work presented at this seminar is related to simplifying and optimizing the production line set up. The first main contribution covers trajectory optimization using variables that can directly be translated to robot code parameters, making it possible to automatically generate robot code. The second main contribution covers robot controller simulations, trying to as closely as possible, but with limited algorithm complexity, approximate robot controllers. This is important when generalizing the problem to different robot vendors and to a larger set of robot commands. The controller trajectory simulation shows good agreement with vendor controllers, and is implemented in the IPS software, where it is denoted Controller Light. It is used during zone size maximization, where smooth and efficient collision free robot trajectories are generated. Furthermore, its accurate representation of tool center point movement is important when performing sealing lay down and spray-painting simulations. The presented contributions have value to automation engineers setting up, or changing, their production line. They cover both simplifications in the workflow for the engineer, as well as improving the final solutions with respect to cycle time and smoothness.

Category Seminar
Location: EF, lecture hall, Hörsalsvägen 11, EDIT trappa C, D och H
Starts: 30 January, 2020, 13:15
Ends: 30 January, 2020, 14:15

Published: Fri 17 Jan 2020.