7 results
Now showing1 - 7 of 7
- Some of the metrics are blocked by yourconsent settings
Item type:Publication, Access status: with files , Data-driven model predictive control with matrix forgetting factorIn this paper, a model predictive control (MPC) based on the Koopman operator framework is presented. This framework, allows to obtain a model solely from data. Moreover, it can be easily updated using a recursive least squares (RLS) approach. This aids the controller to deal with inaccuracies in the model parameters. The performance of the update mechanism can be enhanced by the inclusion of forgetting techniques. One of these is the addition of a constant forgetting factor (CFF) to the update equations. However, this technique can be very sensitive to noise and if the measured signals are not persistently exciting, numerical issues can occur. To avoid the latter, the matrix forgetting factor (MFF) presented in (Bruce et al., 2020) is used. This approach combines two forgetting techniques: variable-rate and variable-direction forgetting, which leads to a higher closed loop performance of the proposed controller.Publicationtype: Conference Paper20 151 - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Access status: with files , LPV system identification and order reduction with application to turbocharged SI enginesThe basis for the model-based design of linear parameter-varying (LPV) controllers is a model of the controlled system. However, models that approximate the system behavior well are often missing in industrial applications. In this thesis, methods for the identification and order reduction of discrete-time LPV models are presented. They are applied to measurements from a real system, the air path of a turbocharged gasoline engine, and show improved computational efficiency and/or provide models of higher accuracy compared to existing methods. Furthermore, the resulting LPV models are suitable for established control design tools.Publicationtype: Doctoral Thesis258 531 - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Access status: without files , Koopman Operator-based model predictive control with recursive online updatePublicationtype: Conference Paper77 - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Access status: without files , Identification of Linear Parameter-Varying Models with Unknown Parameter Dependence Using ϵ-Support Vector Regression(2018-08-09); ; The bias-variance trade-off is very sensitive to the prior selection of functional dependencies when identifying linear parameter-varying (LPV) systems. To counteract this difficulty, various non-parametric methods have been recently proposed. These methods avoid the manual selection of the functional dependency but rather learn it during the identification itself. In this paper, we propose an algorithm to implement the ϵ-tube support vector regression (SVR) approach to such an LPV identification problem. We use some results from the machine learning literature to tune the parameters and provide different directions one could take to optimize these parameters. We demonstrate the effectiveness of our method on an example and compare the results with other methods recently proposed. We observe that because of the insensitive ϵ-tube, the number of parameters was greatly reduced still maintaining the same accuracy in terms of the best fit ratio.Publicationtype: Conference Paper180 - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Access status: without files , LPV state-space identification via IO methods and efficient model order reduction in comparison with subspace methodsIn this paper, we introduce a procedure for global identification of linear parameter-varying (LPV) discrete-time state-space (SS) models with a static, affine dependency structure in a computationally efficient way. The aim is to develop off-the-shelf LPV-SS estimation methods to make identification practically accessible. The benefits of identifying a computational straightforward LPV input-output (IO) model-that has an equivalent SS representation with static, affine dependency-is combined with an LPV-SS model order reduction. To increase practical relevance of the proposed scheme, in this paper, we present a computational attractive model order reduction scheme based on the LPV Ho-Kalman like realization scheme. We analyze the computational complexity and scalability of our method and compare its benefits to the PBSIDopt scheme. Two examples are provided to demonstrate that our introduced approach performs similar to PBSIDopt in a numerical example and outperforms the PBSIDopt on measurements of a real world system, the air-path system of a gasoline engine.Publicationtype: Conference Paper163 - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Access status: without files , Identification of linear parameter-varying systems via IO and subspace identification - A comparisonA common challenge associated with designing a Linear Parameter-Varying (LPV) controller for a nonlinear plant is identification of a plant model in state-space form with static parameter dependency. Till date, the two most frequently used methods to solve this problem are special canonical forms of Input-Output (IO) and subspace (SS) identification. This paper compares both identification methods on various criteria with a focus on low-complexity LPV systems. Features of both methods are illustrated by identification results on a simulated MIMO LPV system taken from the literature and by experimental results on the air-path system of a gasoline engine. Finally, some insight into the selection of an identification method for a given model class is provided.Publicationtype: Conference Paper93 - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Access status: without files , Linear parameter-varying control for air/fuel ratio in SI engines with parameter dependent time delayThis paper is concerned with the air/fuel ratio control of direct injected, spark ignition (SI) combustion engines. The air/fuel ratio is essential for torque generation together with emissions and requires therefore a precise regulation. The system dynamics are modeled for controller design as a first order linear parameter-varying (LPV) system plus parameter varying time delay where the time delay is approximated by a first order Padé function. The scheduling signals determined by the engine operating point are reduced to a single scheduling parameter. The functional dependence on the plant parameters stored in look-up tables is then approximated by rational functions of this scheduling parameter. For controller synthesis a mixed sensitivity design (four-block) is used with parameter dependent weighting to adjust the requirements to the varying plant dynamics. An LPV controller with guaranteed stability and induced L2-norm performance is then obtained by solving linear matrix inequalities (LMIs) on a prespecified grid. Simulation results are presented to validate the designed LPV controller. They show performance superior to that of a scheduled PID controller used in serial production cars.Publicationtype: Conference Paper62