@article{AmaralFiorentinGoellinger2020, author = {Amaral, Jana{\´i}na R. and Fiorentin, Thiago Antonio and G{\"o}llinger, Harald}, title = {Improvement of vehicle stability using a controller based on reinforcement learning}, volume = {14}, journal = {International Journal of Vehicle Systems Modelling and Testing (IJVSMT)}, number = {2-3}, publisher = {Inderscience}, address = {Genf}, issn = {1745-6436}, doi = {https://doi.org/10.1504/IJVSMT.2020.111677}, pages = {97 -- 112}, year = {2020}, language = {en} } @article{FunkDrechslerFiorentinGoellinger2021, author = {Funk Drechsler, Maikol and Fiorentin, Thiago Antonio and G{\"o}llinger, Harald}, title = {Actor-Critic Traction Control Based on Reinforcement Learning with Open-Loop Training}, volume = {2021}, pages = {4641450}, journal = {Modelling and Simulation in Engineering}, publisher = {Hindawi}, address = {New York}, issn = {1687-5605}, doi = {https://doi.org/10.1155/2021/4641450}, year = {2021}, abstract = {The use of actor-critic algorithms can improve the controllers currently implemented in automotive applications. This method combines reinforcement learning (RL) and neural networks to achieve the possibility of controlling nonlinear systems with real-time capabilities. Actor-critic algorithms were already applied with success in different controllers including autonomous driving, antilock braking system (ABS), and electronic stability control (ESC). However, in the current researches, virtual environments are implemented for the training process instead of using real plants to obtain the datasets. This limitation is given by trial and error methods implemented for the training process, which generates considerable risks in case the controller directly acts on the real plant. In this way, the present research proposes and evaluates an open-loop training process, which permits the data acquisition without the control interaction and an open-loop training of the neural networks. The performance of the trained controllers is evaluated by a design of experiments (DOE) to understand how it is affected by the generated dataset. The results present a successful application of open-loop training architecture. The controller can maintain the slip ratio under adequate levels during maneuvers on different floors, including grounds that are not applied during the training process. The actor neural network is also able to identify the different floors and change the acceleration profile according to the characteristics of each ground.}, language = {en} } @inproceedings{PourtoulidouStegerBurgeretal.2019, author = {Pourtoulidou, Despoina and Steger, Fabian and Burger, Uli and G{\"o}llinger, Harald and K{\"o}nig, Ludwig and Obermeier, Erwin and Schweiger, Hans-Georg and Frey, Andreas}, title = {Mission-framed Project-based Learning and Teaching}, booktitle = {Varietas delectat… Complexity is the new normality: SEFI 47th Annual Conference Proceedings}, subtitle = {integrating an Electric Powertrain into a Motor Glider}, editor = {Nagy, Bal{\´a}zs Vince and Murphy, Mike and J{\"a}rvinen, Hannu-Matti and K{\´a}lm{\´a}n, Anik{\´o}}, publisher = {SEFI}, address = {Br{\"u}ssel}, isbn = {978-2-87352-018-2}, doi = {https://www.sefi.be/proceedings/?conference=2019-budapest-isbn-978-2-87352-018-2}, pages = {879 -- 888}, year = {2019}, language = {en} }