Showing papers by "Wpmh Maurice Heemels published in 2015"

Event-Triggered and Self-Triggered Control.

Abstract: Recent developments in computer and communication technologies have led to a new type of large-scale resource-constrained wireless embedded control systems. It is desirable in these systems to limit the sensor and control computation and/or communication to instances when the system needs attention. However, classical sampled-data control is based on performing sensing and actuation periodically rather than when the system needs attention. This article discusses eventand self-triggered control systems where sensing and actuation is performed when needed. Event-triggered control is reactive and generates sensor sampling and control actuation when, for instance, the plant state deviates more than a certain threshold from a desired value. Self-triggered control, on the other hand, is proactive and computes the next sampling or actuation instance ahead of time. The basics of these control strategies are introduced together with references for further reading.

Output-based event-triggered control systems under Denial-of-Service attacks

Abstract: In this paper, we propose a dynamic event-triggered control (ETC) strategy for output-based feedback systems in the presence of Denial-of-Service (DoS) attacks. These malicious DoS attacks aim to impede the communication of measurement data in order to endanger the functionality of the closed-loop system. We show that the proposed ETC scheme, if well designed, can tolerate a class of DoS signals characterized by frequency and duration properties without jeopardizing the stability, performance and Zeno-freeness of the ETC system.

Dynamic event-triggered control under packet losses: The case with acknowledgements

Abstract: In this paper, a dynamic ETC strategy for nonlinear state-feedback systems is proposed that results in guarantees for a finite l p -gain from disturbance input to performance output and a strictly positive lower bound on the inter-event times despite the presence of packet losses. The proposed dynamic ETC strategy has several advantages with respect to the commonly studied static ETC strategy including significantly larger average inter-event times. The proposed design methodology results in tradeoffs between the maximum allowable number of successive packet dropouts, (minimum and average) inter-event times and l p -gains, which will be illustrated by means of a numerical example.

Robust self-triggered model predictive control for constrained discrete-time LTI systems based on homothetic tubes

Abstract: In this paper we present a robust self-triggered model predictive control (MPC) scheme for discrete-time linear time-invariant systems subject to input and state constraints and additive disturbances. In self-triggered model predictive control, at every sampling instant an optimization problem based on the current state of the system is solved in order to determine the input applied to the system until the next sampling instant, as well as the next sampling instant itself. This leads to inter-sampling times that depend on the trajectory of the system. By maximizing the inter-sampling time, the amount of communication in the control system is reduced. In order to guarantee robust constraint satisfaction, Tube MPC methods are employed. Specifically, in order to account for the uncertainty in the system, homothetic sets are used in the prediction of the future evolution of the system. The proposed controller is shown to stabilize a closed and bounded set including the origin in its interior.

Periodic Event-Triggered Control

Abstract: Recent developments in computer and communication technologies are leading to an increasingly networked and wireless world. This raises new challenging questions in the context of networked control systems, especially when the computation, communication and energy resources of the system are limited. To efficiently use the available resources it is desirable to limit the control actions to instances when the system really needs attention. Unfortunately, the classical time-triggered control paradigm is based on performing sensing and actuation actions periodically in time (irrespective of the state of the system) rather than when the system needs attention. Therefore, it is of interest to consider event-triggered control as an alternative paradigm as it is more natural to trigger control actions based on the system state, output, or other available information. Event-triggered control can thus be seen as the introduction of feedback in the sensing, communication, and actuation processes. To facilitate an easy implementation of event-triggered control, we propose to combine the principles and particularly the benefits of event-triggered control and classical periodic time-triggered control. The idea is to periodically evaluate the triggering condition and to decide, at every sampling instant, whether the feedback loop needs to be closed. This leads to the so-called periodic event-triggered control (PETC) systems. In this chapter, we discuss PETC strategies, their benefits and two analysis and design frameworks for linear and nonlinear plants, respectively.

Design of a variable gain integrator with reset

Abstract: This paper studies the properties of a variable gain integrator with reset, i.e. a nonlinear lag filter that is obtained by a) saturating the input, b) filtering the saturated input with a Clegg integrator, and c) add the filtered output to the unsaturated input before applying it to a PID-based controller. Depending on the amount of saturation, the corner frequency of the lag filter is reduced along with the associated phase lag. This follows from a describing function analysis in which at low frequencies a minus 20 dB/decade amplitude decay is realized with a phase lag of only 32.48 degrees. Conditions to assess global asymptotic stability of the closed-loop nonlinear control system are provided that are based on a circle criterion-like argument for the flow condition, which applies to the intervals without resets, combined with a jump condition at reset. The reset integrator design is demonstrated on a piezo-actuated motion system where its favorable phase and amplitude properties induce overshoot and settling times comparable to a single (linear) integrator, but with the disturbance rejection properties of a double integrator.

Resource-aware set-valued estimation for discrete-time linear systems

Abstract: In this paper, we propose a self-triggered estimator for discrete-time linear systems subject to unknown but bounded disturbances affecting both the system states and outputs. The proposed self-triggered estimator is a set-valued estimator that employs rollout techniques to reduce the communication between the sensors and the estimator with respect to a periodic sampling strategy. Moreover, at each time instant it provides an estimate of the plant state and a guaranteed bound on the difference between the true plant state and the estimate.

Feedforward for multi-rate motion control: Enhanced performance and cost-effectiveness

Abstract: In traditional feedback control, a single sampling rate is used for all control loops Consequently, achieving higher performance by increasing the sampling rate is generally costly The aim of this paper is to develop a multi-rate control framework to create a breakthrough in the performance/cost trade-off in digital controller implementation In the proposed approach one of the control loops is implemented at a lower rate of which the feedforward controller is designed through norm-based minimalization of the tracking error in this multi-rate framework By designing and implementing one of the control loops at a lower rate, the cost is reduced and the multi-rate problem is addressed Through simulation the adequate performance of the proposed multi-rate approach is demonstrated

An event-triggered policy for remote sensing and control with performance guarantees

Abstract: We consider a networked control system in which a remote controller queries the plant's sensors for measurement data and decides when to transmit control inputs to the plant's actuators. The goal is to keep transmissions to a minimum while guaranteeing that the closed-loop performance is within acceptable bounds. Our approach extends a recent line of research where explicit event-triggered control policies with performance guarantees are derived using approximate dynamic programming. The proposed policy in this new setting can be separated into an offline scheme for sensor query in which the transmission instances are computed a priori and an online scheme to schedule control input transmissions. The usefulness of the results is illustrated through two numerical examples.

Definitions of incremental stability for hybrid systems

Abstract: The analysis of incremental stability properties typically involves measuring the distance between any pair of solutions of a given dynamical system, corresponding to different initial conditions, at the same time instant. This approach is not directly applicable for hybrid systems in general. Indeed, hybrid systems generate solutions that are defined with respect to hybrid times, which consist of both the continuous time elapsed and the discrete time, that is the number of jumps the solution has experienced. Two solutions of a hybrid system do not a priori have the same time domain, and we may therefore not be able to compare them at the same hybrid time instant. To overcome this issue, we invoke graphical closeness concepts. We present definitions for incremental stability depending on whether incremental asymptotic stability properties hold with respect to the hybrid time, the continuous time, or the discrete time, respectively. Examples are provided throughout the paper to illustrate these definitions, and the relations between these three incremental stability notions are investigated. The definitions are shown to be consistent with those available in the literature for continuous-time and discrete-time systems.

Rollout strategies for output-based event-triggered control

Abstract: Recent research advocates that replacing the periodic communication paradigm by an event-triggered paradigm can have significant benefits for control systems. While in many event-triggered control solutions transmission decisions are based on full state information, in most applications only partial state information is available for feedback (output measurements). Here we propose an optimization-based output-feedback event-triggered solution for linear discrete-time systems which guarantees a performance bound with respect to periodic control, while reducing the communication load. Performance is measured by a quadratic cost. The usefulness of the results is illustrated through a numerical example.

Switching data-processing methods for feedback control: Breaking the speed versus accuracy trade-off

Abstract: In many digitally controlled applications, such as vision-based and data-intensive control, the choice of the data-processing method for distilling (control-relevant) state information is non-trivial and important. While accurate processing methods require significant processing time, limiting the closed-loop control rate, faster methods introduce larger inaccuracies in the distilled state information. This leads to a trade-off between speed and accuracy, which can be considered off-line or on-line. In this paper, we propose an on-line strategy to switch between different data-processing methods in real-time in order to improve closed-loop performance for linear systems.

A lifting approach to ℒ2-gain analysis of periodic event-triggered and switching sampled-data control systems

Abstract: In this work we are interested in the stability and ℒ2-gain of hybrid systems with linear flow dynamics, periodic time-triggered jumps and nonlinear possibly set-valued jump maps. This class of hybrid systems includes various interesting applications such as periodic event-triggered control. In this paper we also show that sampled-data systems with arbitrarily switching controllers can be captured in this framework by requiring the jump map to be set-valued. We provide novel conditions for the internal stability and ℒ2-gain analysis of these systems adopting a lifting-based approach. In particular, we establish that the internal stability and contractivity in terms of an ℒ2-gain smaller than 1 are equivalent to the internal stability and contractivity of a particular discrete-time set-valued nonlinear system. Despite earlier works in this direction, these novel characterisations are the first necessary and sufficient conditions for the stability and the contractivity of this class of hybrid systems. The results are illustrated through multiple new examples.

Model-based reconstruction and feedback control of the plasma particle density in tokamaks

Abstract: Reference EPFL-CONF-218577 URL: http://www.ipfn.ist.utl.pt/EPS2015/ URL: https://crpplocal.epfl.ch/pinboard/papers/154801704.pdf Record created on 2016-05-28, modified on 2017-05-12

Modeling, observer design and robust control of the particle density profile in tokamak plasmas

Abstract: A new approach to real-time estimation and feedback control of the particle density profile in tokamak plasmas is presented, based on ideas from Kalman filtering and H∞ robust control synthesis. Traditionally, the density profile is reconstructed in real-time by solving an inversion problem using a measurement from a single time instant. Such an approach is sensitive to sensor errors and does not account for the dynamical evolution and spatial continuity of the density. The observer-based approach we presented here includes the system dynamics, which is realized by careful modeling of the particle density behaviour using a 1D PDE with a nonlinear source term and two ODEs, which are discretized in space and time to yield a finite-dimensional nonlinear model. The influence of other plasma quantities and operational modes on the transport dynamics are included in the control-oriented model as time-varying parameters. An extended Kalman filter estimates the density, additive random-walk state disturbances as well as fringe jumps (a specific type of sensor error) from measurements, for which special measures are needed. Offline reconstruction using tokamak measurements show accurate estimation of the density profile and show the quality of fringe jump detection. Moreover, a robust state feedback controller with anti-windup is designed based on the model to track a reference signal for the average density, with the estimate obtained from the observer. Closed-loop simulations show that the controller is able to track representative reference signals, with the performance mostly limited by the nonnegativity constraint of the control input.

Frequency-domain analysis of real-time and networked control systems with stochastic delays and data drops

Abstract: We present a novel frequency-domain analysis framework for a closed-loop model capturing a wide range of real-time and networked control systems with stochastic delays and packet drops. Our results allow for inferring the mean and variance of the output response to deterministic inputs, based on a new frequency response plot. We illustrate the usefulness of our results in the context of real-time control systems with input-to-output delays resulting from the use of a shared processor.

Resource-aware control and dynamic scheduling in CPS

Abstract: Recent developments in computer and communication technologies are leading to an increasingly networked and wireless world. This raises new challenging questions in the context of control for cyberphysical systems (CPS), especially when the computation, communication, energy and actuation resources (for control) of the system are limited and/or shared by multiple control tasks. These limitations obstruct the use of classical design techniques for feedback control algorithms and call for new resource-aware control paradigms. These new resource-aware control systems typically have to take both discrete decisions (which task is allowed to use the resource) and continuous decisions (which continuous control input is generated for the task). In this talk two approaches are presented to address this hybrid co-design problem. Both approaches result in control algorithms that exploit real-time measurement information available on the state of the CPS and decide dynamically on the actions to take. This leads to the situation that individual control tasks are no longer executed in classical periodic time-triggered patterns, but in aperiodic patterns with varying inter-execution times. By abandoning the periodic scheduling of control tasks, the aim is to realise better trade-offs between the overall performance of the CPS and the required resource utilisation. The approaches are illustrated by various applications. interesting challenges for the future are discussed as well.

On observability in networked control systems with packet losses

Abstract: This paper deals with observability properties of networked control systems subject to packet losses. We employ a switching systems perspective in which available information on the packet loss signal, e.g., there can be at most a pre-specified number of consecutive losses, is modelled through an automaton. Based on this perspective we address several natural extensions envisioned in [7]. Our specific contributions are as follows. Firstly, we show that the method introduced in [7] in the context of controllability of linear systems subject to packet losses extends to the question of observability. The proposed characterization is necessary and sufficient as well as algorithmically verifiable. For the observability problem, our proof is valid also for non-invertible matrices, thereby improving upon the previous results in [7]. Secondly, we show that the model employed for our analysis encompasses the model of wireless control networks with switching delays introduced in [6] (though at a cost of exponential encoding). We raise several open questions related to the algebraic nature of the problem under consideration.

Resource-Aware Control and Dynamic Scheduling in CPS

Abstract: Recent developments in computer and communication technologies are leading to an increasingly networked and wireless world. This raises new challenging questions in the context of control for cyber-physical systems (CPS), especially when the computation, communication, energy and actuation resources (for control) of the system are limited and/or shared by multiple control tasks. These limitations obstruct the use of classical design techniques for feedback control algorithms and call for new resource-aware control paradigms. These new resource-aware control systems typically have to take both discrete decisions (which task is allowed to use the resource) and continuous decisions (which continuous control input is generated for the task). In this talk two approaches are presented to address this hybrid co-design problem. Both approaches result in control algorithms that exploit real-time measurement information available on the state of the CPS and decide dynamically on the actions to take. This leads to the situation that individual control tasks are no longer executed in classical periodic time-triggered patterns, but in aperiodic patterns with varying inter-execution times. By abandoning the periodic scheduling of control tasks, the aim is to realise better trade-offs between the overall performance of the CPS and the required resource utilisation. The approaches are illustrated by various applications. interesting challenges for the future are discussed as well.

RobustSelf-TriggeredMPCforConstrainedLinearSystems: ATube-BasedApproach ?

Abstract: We propose a robust self-triggered control algorithm for constrained linear discrete-time systems subject to additive disturbances based on MPC. At every sampling instant, the controller provides both the next sampling instant, as well as the inputs that are applied to the system until the next sampling instant. By maximizing the inter-sampling time subject to bounds on the MPC value function, the average sampling frequency in the closed-loop system is decreased while guaranteeing an upper bound on the performance loss when compared with an MPC scheme sampling at every point in time. Robust constraint satisfaction is achieved by tightening input and state constraints based on a Tube MPC approach. Moreover, a compact set in the state space, which is a parameter in the MPC scheme, is shown to be robustly asymptotically stabilized.