Deep reinforcement learning for gearshift controllers in automatic transmissions

Gaiselmann, Gerd; Altenburg, Stefan; Studer, Stefan; Peters, Steven

doi:10.1016/j.array.2022.100235

Cited by 7 publications

(3 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It allows to expand the functionality of the motor unit car by multidirectional reduction of power coming from the engine to the transmission system, gear ratio control, increasing the operating range of speed and torque, providing idling and reversing the motion of the vehicle (Fig. 1) [31][32][33][34][35].…”

Section: Resultsmentioning

confidence: 99%

See 1 more Smart Citation

Automating the Assembly Process of Passenger Car Gearboxes

Filatov¹

2022

AJRCoS

View full text Add to dashboard Cite

Aims: The study aims to identify ways and potential solutions to automate the assembly and production process for passenger car gearboxes. Object of Research: Assembly and production process for car gearboxes. Subject of Research: Modern and evolutionary automation tools that have the potential to be implemented in the assembly processes of automotive transmission controls. Methodology: To achieve this goal, as part of this study, it is planned to apply methods of bibliometric analysis of leading scientometric databases to obtain correlation relationships and analytical conclusions regarding the vector of development of automation means of passenger car gearbox assembly process. Results: As a result of the research by means of scientometric analysis and correlation the vector of probabilistic technical solutions of integration and development of automation means of the sequence of production operations during the assembly of the transmission, as well as adaptive framework-design solutions for the implementation of tools of the fourth iteration of industrial-industrial progress in the production processes of assembly of the studied technical control means and logical-technological connection of the elements of the transmission system, which affect the overall process of automotive manufacturization. Conclusion: The passenger car gearbox is a multi-component, complex system whose assembly is a complex multi-operational process, and given the high responsibility of this machine element, there is an urgent need to introduce modern automation tools into the assembly and production processes, which will significantly optimize global automatofactoring. The practical results of the present study consist in the formation of a focus scientometric database of profile data, identification of a potential vector of development of means and systems of automation of assembly-production operations, identification and formation of solutions for the implementation of modern means of automated production in the actual global automotive manufacturing, which allows to get the optimum ratio of production costs/quality of products by improving the manufacturability, productivity and flexibility of processes of assembly of multi-element and multi-component automotive systems and structures.

show abstract

Section: Resultsmentioning

confidence: 99%

“…For the studied element of the automobile transmission the gearbox currently uses methods of semi-automatic machine-human assembly [31][32][33][34][35]49,50] (Fig. 7).…”

Section: Resultsmentioning

confidence: 99%

Automating the Assembly Process of Passenger Car Gearboxes

Filatov¹

2022

AJRCoS

View full text Add to dashboard Cite

show abstract

“…The development of neural networks and reinforcement learning in recent years has provided additional options for designing feedback controllers [16]. Many researchers have explored the use of these techniques for various aspects of car shifting, such as those in references [17]- [19]. Reinforcement learning algorithms can be divided into model-based and modelfree categories depending on whether they use a model.…”

Section: Introductionmentioning

confidence: 99%

The Optimization of RBFNN Gearshift Controller Parameters for Electric Vehicles Using PILCO Reinforcement Learning

Liu,

Zhang,

et al. 2023

IEEE Access

View full text Add to dashboard Cite

In order to improve the efficiency of gearshift controller parameter optimization and obtain a good gearshift controller control effect, this article proposes an optimization method for electric vehicle transmission gearshift controller, and selects dual clutch transmission as the research object that establishes an 11-degree-of-freedom gearshift dynamics model and a feedforward-feedback gearshift control model. The feedforward control is chosen from the target trajectory given by the Legendre pseudo-spectral approach, and the feedback controller is a Gaussian kernel radial basis function neural network controller. The feedback controller performs parameter optimization by the Probabilistic Inference for Learning Control (PILCO) reinforcement learning algorithm to obtain a control strategy that matches the actual gearshift conditions. By comparing how well the master-slave moving disk can follow the target trajectory during various optimization iterations, it is verified that the algorithm requires only a few experiments to complete the optimization, and the optimized Radial Basis Function Neural Network (RBFNN) control has a better control effect by comparing the results of different iterations. Applying the learned controller to various slope and load circumstances yields data that demonstrate that all can have an obvious optimization effect with good robustness. Additionally, the reinforcement learning technique suggested in this research can be used for various gearshift controller parameter optimization to assist engineers and technicians in increasing the effectiveness of their Research and Development.INDEX TERMS Electric vehicle, dual clutch transmission, gearshift control, reinforcement learning.

show abstract

Nutzung von KI-Methoden für die Kupplungsentwicklung in automobilen Antriebssträngen

Altenburg

Schuchter

Bause

et al. 2023

Forsch Ingenieurwes

View full text Add to dashboard Cite

ZusammenfassungIm Spannungsfeld steigender Erwartungen an Fahrkomfort und Energieeffizienz stoßen aktuelle Entwicklungsmethoden der Modellbildung und Optimierung für Fahrzeugkupplungen an ihre Grenzen. In diesem Beitrag wird der Einsatz von KI-Methoden für die Kupplungsentwicklung untersucht und ein Überblick anhand verschiedener Anwendungsbeispiele in aktuellen Forschungsprojekten der Mercedes-Benz AG gegeben. Mittels überwachten Lernens und tiefen neuronalen Netzen werden ein Reibungszahlmodell sowie ein Temperaturmodell einer Kupplung mit hoher Abbildungsgüte entwickelt. Verstärkendes Lernen mit tiefen neuronalen Netzen wird zur Synthese von Regelungen für verschiedene Gangwechsel eingesetzt. Fahrzeugmessdaten werden mit Cluster-Algorithmen analysiert, um Handlungsempfehlungen für die Applikation des Motorwiederstarts eines hybriden Antriebsstrangs abzuleiten. Mit den gezeigten Methoden steigt das Automatisierungspotential in der Entwicklung und der Aufwand für die Übernahme komplexer Entwicklungsprozesse auf neue Getriebevarianten sinkt.

show abstract

Deep reinforcement learning for gearshift controllers in automatic transmissions

Cited by 7 publications

References 20 publications

Automating the Assembly Process of Passenger Car Gearboxes

Automating the Assembly Process of Passenger Car Gearboxes

The Optimization of RBFNN Gearshift Controller Parameters for Electric Vehicles Using PILCO Reinforcement Learning

Nutzung von KI-Methoden für die Kupplungsentwicklung in automobilen Antriebssträngen

Contact Info

Product

Resources

About