Reinforcement learning in real-time geometry assurance
Paper i proceeding, 2018

To improve the assembly quality during production, expert systems are often used. These experts typically use a system model as a basis for identifying improvements. However, since a model uses approximate dynamics or imperfect parameters, the expert advice is bound to be biased. This paper presents a reinforcement learning agent that can identify and limit systematic errors of an expert systems used for geometry assurance. By observing the resulting assembly quality over time, and understanding how different decisions affect the quality, the agent learns when and how to override the biased advice from the expert software.

expert advice

reinforcement learning

geometry assurance

Författare

Emilio Jorge

Stiftelsen Fraunhofer-Chalmers Centrum för Industrimatematik

Lucas Brynte

Stiftelsen Fraunhofer-Chalmers Centrum för Industrimatematik

Constantin Cronrath

Chalmers, Elektroteknik, System- och reglerteknik

Oskar Wigström

Chalmers, Elektroteknik, System- och reglerteknik

Kristofer Bengtsson

Chalmers, Elektroteknik, System- och reglerteknik

Emil Gustavsson

Stiftelsen Fraunhofer-Chalmers Centrum för Industrimatematik

Bengt Lennartson

Chalmers, Elektroteknik, System- och reglerteknik

Mats Jirstrand

Stiftelsen Fraunhofer-Chalmers Centrum för Industrimatematik

Procedia CIRP

22128271 (eISSN)

Vol. 72 1073-1078

51st CIRP Conference on Manufacturing Systems, CIRP CMS 2018
Stockholm, Sweden,

Smart Assembly 4.0

Stiftelsen för Strategisk forskning (SSF) (RIT15-0025), 2016-05-01 -- 2021-06-30.

Ämneskategorier

Programvaruteknik

Datavetenskap (datalogi)

Datorsystem

DOI

10.1016/j.procir.2018.03.168

Mer information

Senast uppdaterat

2023-01-20