8 results
Now showing1 - 8 of 8
- Some of the metrics are blocked by yourconsent settings
Item type:Publication, Access status: without files , Distributed Controller Design for Systems Interconnected over Chordal Graphs(2020-07); ; In this paper we consider the problem of designing distributed H2 suboptimal controllers for continuoustime networked systems in a distributed fashion, when both the plant and the controller are interconnected over the same chordal network. Combining results for networked systems and sparse semidefinite optimization, it is possible to reduce the conservatism introduced by the decomposition of the synthesis problem compared to existing methods, while at the same time also reducing the computational complexity.Publicationtype: Conference Paper113 - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Access status: without files , Distributed Control of Large Scale Interconnected Systems with Boundary Conditions-Shift Operator vs. Interconnection Graph(2020-05); ; Different modeling and control strategies for distributed systems described by partial differential equations, where the involved signals depend not only on a temporal but also on one or more spatial variables, are investigated. When applying an array of actuator/sensor pairs, a spatial discretization is induced and - assuming spatial invariance - the system can be modeled as an interconnection of identical subsystems together with boundary conditions. While centralized control methods quickly turn impractical for such systems due to the high number of input and output channels, low-complexity synthesis techniques based on either a spatial shift operator or on interconnection graphs offer a solution for this problem at the cost of some conservatism. In order to compare the methods with respect to performance and complexity in terms of computation and implementation, the simple model of the temperature distribution in a thin rod with two different boundary conditions is used. Three types of distributed controllers are compared to an optimal centralized output feedback controller, obtained using standard H∞-loopshaping.Publicationtype: Conference Paper62 - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Access status: with files , Spatio-temporal loop shaping for distributed control of PDE systemsThe systems of interest in this paper are described by a possibly large number of interconnected subsystems, where the spatial discretization into subsystems is induced by applying an array of collocated actuator/sensor pairs. When considering these types of systems in a classical centralized fashion, they have a high number of input/output signals on one hand, and usually a sparse structure on the other. While the method of loop-shaping in classical control is well-established, the notion of spatially distributed systems makes it possible to extend frequency domain loop shaping to spatial frequencies and to include weighting filters with spatio-temporal dynamics. By using a Fourier transform, controller synthesis can be done for each spatial frequency at a time. This method is applied to the control of the temperature distribution of a thin metal rod.Publicationtype: Conference Paper89 195 - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Access status: without files , Distributed Modeling of Structural Systems Based on Finite Element Methods with Application to an Actuated Beam(2019-12); ; Publicationtype: Conference Paper153 - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Access status: without files , A Decomposition Approach for a Class of Spatially Interconnected Systems of Finite Spatial Extent(2018-08-09); ; This paper presents a decomposition approach for a class of spatially interconnected systems of finite extent obtained by discretizing partial differential equations including boundary conditions. Such finite extent systems can be modeled as an interconnection of identical subsystems communicating with their nearest neighbors resulting in a lattice-interconnection-topology. Recently the question has been raised whether this class of systems is decomposable. It is shown that these systems are decomposable with respect to a pattern matrix for which a construction law is presented, using the coefficients of the finite difference equation and taking the boundary conditions into account. By bringing the system into a decomposed form, several existing distributed analysis and synthesis techniques can be applied. This approached also helps to bridge the gap between spatially interconnected and multiagent systems. Finally, the approach is demonstrated with stability analysis on two numerical examples.Publicationtype: Conference Paper152 - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Access status: without files , Well-posedness & stability analysis for spatially interconnected systems with finite extentNovel sufficient conditions for well-posedness and stability of spatially interconnected systems with finite extent are proposed. Existing conditions for space-invariant interconnected systems do in general not guarantee these properties for finite-extent systems. Rather than using a model based on the spatial shift operator, here the system is represented as a network of (not necessarily identical) subsystems with a particular interconnection topology. The analysis problem is solved by utilizing the full-block S-procedure; by imposing a suitable structure on the associated multiplier the problem can be reduced to a complexity that does not depend on the size of the network but only on the size of a single subsystem and the number of different subsystem types present in the network. Various boundary conditions can be taken into account. The proposed method is illustrated on the model of a thermal rod equipped with an array of sensors and actuators.Publicationtype: Conference Paper99 - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Access status: without files , Active vibration control of an aluminum beam - An experimental testbed for distributed vs. centralized control(2017); This paper presents an experimental test bench for assessing the practicality of distributed control strategies in the framework of spatially-interconnected systems. To compare low-complexity distributed control schemes with multi-input multioutput (MIMO) control and to assess the results experimentally requires a test bed that is sufficiently complex to display on one hand the characteristic features of a distributed system, and yet is not too complex in order not to render the MIMO synthesis problem intractable. Here a 5m long aluminum beam, equipped with 16 collocated pairs of piezoelectric actuators and sensors, is used for vibration control experiments. A MIMO model obtained by finite element modeling techniques is presented and a distributed model in the spatially-interconnected systems framework is derived. Model reduction techniques are applied to the MIMO model and a centralized controller is designed to attenuate the vibration using H∞-loopshaping techniques. Finally, the performance of the controller is experimentally validated on the beam.Publicationtype: Conference Paper156 - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Access status: without files , A decentralized asymmetric weighting approach for improved convergence of multi-agent systems with undirected interactionThis work considers the convergence rate of multi-agent systems with discrete-time single-integrator dynamics and undirected interaction topologies. In recent work it has been proven that in case of lattice interaction topologies the convergence rate can be bounded away from zero, independent of the network size, using asymmetric weightings that give the interaction graph a preferred communication direction. Approximation methods for more general graphs, based on relative angles between agents, are presented, which suggest that the convergence rate is bounded away from zero as well. This work proposes alternative approximation methods, that improve the convergence rate compared to previous approximation methods. Furthermore it is shown that the improvement of the approximation methods degrade in comparison to other weighting approaches, the more the considered topology differs from a lattice graph. Therefore an iterative algorithm is proposed, that extends the idea of a preferred communication direction to general graphs, which are not similar to lattices or where the relative angles are not known.Publicationtype: Conference Paper28