Reinforcement Learning Framework for the self-learning Suppression of Clutch Judder in automotive Drive Trains
Abstract:
In electromechanically actuated clutches, the active damping of vibrations by means of control of the clamping force allow the use of high performance materials in the friction pairing, which makes a more energy and cost efficient design of the clutch. In this work, a reinforcement learning framework for the control of the clamping force for the active suppression of judder vibrations is proposed and developed.
| Zugehörige Institution(en) am KIT | Institut für Produktentwicklung (IPEK) |
| Publikationstyp | Hochschulschrift |
| Publikationsjahr | 2016 |
| Sprache | Englisch |
| Identifikator | urn:nbn:de:swb:90-614367 KITopen-ID: 1000061436 |
| Verlag | Karlsruher Institut für Technologie (KIT) |
| Umfang | 170 S. |
| Art der Arbeit | Dissertation |
| Fakultät | Fakultät für Maschinenbau (MACH) |
| Institut | Institut für Produktentwicklung (IPEK) |
| Prüfungsdaten | 26.01.2016 |
| Schlagwörter | Reinforcement Learning, Kupplungsrupfen |
| Referent/Betreuer | Albers, A. |