Stochastic EventBased Control and Estimation
(2012) In PhD Theses TFRT1095. Abstract
 Digital controllers are traditionally implemented using periodic sampling, computation, and actuation events. As more control systems are implemented to share limited network and CPU bandwidth with other tasks, it is becoming increasingly attractive to use some form of eventbased control instead, where precious events are used only when needed.
Forms of eventbased control have been used in practice for a very long time, but mostly in an adhoc way. Though optimal solutions to most eventbased control problems are unknown, it should still be viable to compare performance between suggested approaches in a reasonable manner.
This thesis investigates an eventbased variation on the stochastic... (More)  Digital controllers are traditionally implemented using periodic sampling, computation, and actuation events. As more control systems are implemented to share limited network and CPU bandwidth with other tasks, it is becoming increasingly attractive to use some form of eventbased control instead, where precious events are used only when needed.
Forms of eventbased control have been used in practice for a very long time, but mostly in an adhoc way. Though optimal solutions to most eventbased control problems are unknown, it should still be viable to compare performance between suggested approaches in a reasonable manner.
This thesis investigates an eventbased variation on the stochastic linearquadratic (LQ) control problem, with a fixed cost per control event. The sporadic constraint of an enforced minimum interevent time is introduced, yielding a mixed continuous/discretetime formulation. The quantitative tradeoff between event rate and control performance is compared between periodic and sporadic control. Example problems for firstorder plants are investigated, for a single control loop and for multiple loops closed over a shared medium.
Path constraints are introduced to model and analyze higherorder eventbased control systems. This componentbased approach to stochastic hybrid systems allows to express continuous and discretetime dynamics, state and switching constraints, control laws, and stochastic disturbances in the same model. Sumofsquares techniques are then used to find bounds on control objectives using convex semidefinite programming.
The thesis also considers state estimation for discrete time linear stochastic systems from measurements with convex set uncertainty. The Bayesian observer is considered given logconcave process disturbances and measurement likelihoods. Strong logconcavity is introduced, and it is shown that the observer preserves logconcavity, and propagates strong logconcavity like inverse covariance in a Kalman filter. A recursive state estimator is developed for systems with both stochastic and setbounded process and measurement noise terms. A timevarying linear filter gain is optimized using convex semidefinite programming and ellipsoidal overapproximation, given a relative weight on the two kinds of error. (Less)
Please use this url to cite or link to this publication:
http://lup.lub.lu.se/record/3166619
 author
 Henningsson, Toivo ^{LU}
 supervisor

 Anton Cervin ^{LU}
 opponent

 Professor Hespanha, Joao Pedro, University of California, Santa Barbara, USA
 organization
 publishing date
 2012
 type
 Thesis
 publication status
 published
 subject
 keywords
 eventbased control, eventbased estimation, sporadic control, stochastic control, stochastic hybrid systems, sumofsquares methods, control over networks, quantized measurements
 in
 PhD Theses
 volume
 TFRT1095
 pages
 183 pages
 publisher
 Department of Automatic Control, Lund Institute of Technology, Lund University
 defense location
 Lecture hall M:B, Mbuilding, Ole Römers väg 1, Lund University Faculty of Engineering
 defense date
 20121218 10:15
 ISSN
 02805316
 ISBN
 9789174734102
 language
 English
 LU publication?
 yes
 id
 de922e8b95044ddabaff94bdf6b53237 (old id 3166619)
 date added to LUP
 20121122 16:05:51
 date last changed
 20160919 08:44:48
@phdthesis{de922e8b95044ddabaff94bdf6b53237, abstract = {Digital controllers are traditionally implemented using periodic sampling, computation, and actuation events. As more control systems are implemented to share limited network and CPU bandwidth with other tasks, it is becoming increasingly attractive to use some form of eventbased control instead, where precious events are used only when needed.<br/><br> <br/><br> Forms of eventbased control have been used in practice for a very long time, but mostly in an adhoc way. Though optimal solutions to most eventbased control problems are unknown, it should still be viable to compare performance between suggested approaches in a reasonable manner.<br/><br> <br/><br> This thesis investigates an eventbased variation on the stochastic linearquadratic (LQ) control problem, with a fixed cost per control event. The sporadic constraint of an enforced minimum interevent time is introduced, yielding a mixed continuous/discretetime formulation. The quantitative tradeoff between event rate and control performance is compared between periodic and sporadic control. Example problems for firstorder plants are investigated, for a single control loop and for multiple loops closed over a shared medium.<br/><br> <br/><br> Path constraints are introduced to model and analyze higherorder eventbased control systems. This componentbased approach to stochastic hybrid systems allows to express continuous and discretetime dynamics, state and switching constraints, control laws, and stochastic disturbances in the same model. Sumofsquares techniques are then used to find bounds on control objectives using convex semidefinite programming.<br/><br> <br/><br> The thesis also considers state estimation for discrete time linear stochastic systems from measurements with convex set uncertainty. The Bayesian observer is considered given logconcave process disturbances and measurement likelihoods. Strong logconcavity is introduced, and it is shown that the observer preserves logconcavity, and propagates strong logconcavity like inverse covariance in a Kalman filter. A recursive state estimator is developed for systems with both stochastic and setbounded process and measurement noise terms. A timevarying linear filter gain is optimized using convex semidefinite programming and ellipsoidal overapproximation, given a relative weight on the two kinds of error.}, author = {Henningsson, Toivo}, isbn = {9789174734102}, issn = {02805316}, keyword = {eventbased control,eventbased estimation,sporadic control,stochastic control,stochastic hybrid systems,sumofsquares methods,control over networks,quantized measurements}, language = {eng}, pages = {183}, publisher = {Department of Automatic Control, Lund Institute of Technology, Lund University}, school = {Lund University}, series = {PhD Theses}, title = {Stochastic EventBased Control and Estimation}, volume = {TFRT1095}, year = {2012}, }