Emeritus Laureate Professor Graham Goodwin

This chapter further explores the connection between continuous-time power spectral densities and their sampled data equivalents. A key issue of importance is the existence of sampling zeros in the sampled-data power spectral density. These can be asymptotically characterized as the sampling period goes to zero. The results are the stochastic extension of the results presented in Chap. 5 for the deterministic case.

This chapter develops various approximate discrete-time models for general linear deterministic systems. It is always possible to obtain an exact sampled models for linear systems. However, approximate models are treated here to provide insights into the structure of discrete-time models, to obtain simpler models, and to be able to construct similar approximate sampled models for nonlinear systems, later in the book. We present models based on simple Euler integration, on the inclusion of asymptotic sampling zeros, on up-sampling, on normal forms and on truncated Taylor series expansions.

In this chapter, background on sampling and Fourier analysis of signals is provided. The chapter provides a brief review of the key concepts to establish notation. In particular, the continuous and discrete-time Fourier transform are defined. Sampling of signals and the aliasing effect are then presented. Finally, the use of antialiasing filters and reconstruction of signals from samples are also discussed.

This chapter extends the ideas presented in Chap. 8 to develop discrete-time models for nonlinear deterministic systems. In particular, we develop sampled models based on up-sampling and on truncated Taylor series expansion of the normal form. The approximation errors are precisely characterized.

The zeros of the sampled model depend on the artifacts of the sampling process, namely, how the continuous-time input is generated by a hold device and how the output samples are obtained. In this chapter, the effect of the hold device is studied in a more general setting. In particular, hold devices are characterized by their impulse response. In this framework, zero- and first-order holds are particular cases of these generalized hold functions.

This chapter extends the review of deterministic signals presented in Chap. 2 to cover sampling and Fourier analysis of stochastic signals. A brief summary of key concepts, such as continuous-time and sampled power spectral densities, is provided to establish notation and core concepts.

This chapter considers the main topic of interest in this book, namely: how to obtain a sampled-data model for a given continuous-time system. Sampled models for linear systems with deterministic inputs are first considered, for which exact sampled-data model can be obtained. Unless stated explicitly otherwise, it is assumed that the anti-aliasing filter, if present, is included as part of the continuous-time system model.

This chapter explores the relationship between sampled-data models and the associated continuous-time system. In particular, the chapter shows how the sampled-data model relates to the original continuous-time model, when a fast sampling rate is used. It is shown that a simple reparameterization of the sampled model in the so called incremental (or delta) form, is required in order to recover the continuous-time model when the sampling period goes to zero.

In this chapter the ideas in Chap. 16 are extended to obtain approximate sampled models for stochastic nonlinear systems. In particular, the concepts of time-domain approximate stochastic sampled models based on up-sampling and successive integration are extended to the nonlinear case.

This chapter investigates the application of the approximate sampled-data models to problems arising in estimation and control, both for linear and nonlinear systems. Applications include state estimation and feedback control for linear systems, sampled-data state feedback control for nonlinear systems, continuous-time system identification from sampled data, and sampled-data predictive control of an electrical machine.

Most of the book deals with sampled models that describe the signals at the sampling instants. In this chapter we focus on describing the system response that occurs between samples. The ideas presented here build on the results for the at-sample response previously presented in the book. Two views of the intersample response are presented, namely in the frequency domain and the time domain.

This chapter extends the notion of incremental models (presented in Chap. 4) to the stochastic case. In particular, we discuss the advantages of representing in incremental form the sampled-data models for stochastic systems so as to obtain a well defined limit as the sampling period approaches zero.

This chapter explores several applications of approximate stochastic sampled-data models. It is shown that, in some applications, an understanding of the role of sampling zero dynamics can be crucial in obtaining accurate results. In particular, parameter and state estimation problems are considered where the use of a restricted bandwidth may be crucial to correctly account for the artifacts of sampling.

In this chapter we extend the ideas of Chap. 16 to the nonlinear case. We first describe aspects of nonlinear stochastic models based on stochastic calculus. Important technical issues arise such as the Ito rule and Ito-Taylor expansions for stochastic processes. These ideas are applied to numerical solutions of stochastic differential equations.

This chapter shows how approximate sampled-data models for stochastic-linear systems can be developed. In the linear case exact sampled-data models can always be obtained. However, results on approximate linear-stochastic sampled-data models are developed here, as a prelude to the nonlinear case treated in the next chapter.

The location of poles of sampled models depends on the poles of the underlying continuous-time system and the sampling period. However, the relation between the zeros in discrete-time and continuous-time is much more involved. The sampled model can have extra zeros which have no continuous-time counterpart. This chapter examines the asymptotic behaviour of these sampling zeros in the discrete-time model, as the sampling period goes to zero.

The Euler¿Frobenius polynomials have been throughout the book to characterize asymptotic sampling zeros in discrete models as the sampling period goes to zero. This chapter presents a brief historical account of these polynomials, and a summary of (equivalent) definitions and properties found in the literature.

This chapter explores the robustness issues associated with the use of sampled models. The effect of folding of high frequency aspects of the continuous-time model back to lower frequencies is discussed. This aliasing effect may impact the assumptions regarding the high frequency behaviour of the system due to unmodeled high frequency poles or zeros. In particular, the location of asymptotic sampling zeros rely on these high frequency characteristics. Thus, one needs to be careful about the bandwidth of validity of models.

In this chapter, the impact of the choice of anti-aliasing filter on the resultant stochastic sampled-data model is explored. The zeros of the sampled output power spectral density depend on the choice of this filter. In particular, we present a generalized sampling filter design, such that the sampling zeros of the sampled output power spectral density are asymptotically assigned. This analysis is the stochastic counterpart of generalized hold functions for deterministic models, presented in Chap. 6.

In this chapter we review sampling issues for stochastic signals. These ideas are used to formulate models for stochastic linear systems. The impact of the antialiasing filter used prior to sampling is also discussed.

Fractional-order systems have received considerable attention by the engineering community in the last years. This chapter extends the ideas presented in Chaps. 5 and 8 to systems having fractional-order dynamics. Specifically, fractional-order Euler¿Frobenius polynomials are defined and used to characterize the asymptotic sampling zeros for fractional systems as the sampling period tends to zero.

Whilst most physical systems occur naturally in continuous time, it is necessary to deal with sampled data for identification purposes. In principle, one can derive an exact sampled data model for any given linear system by integration. However, conversion to sampled data form implicitly involves folding of high-frequency system characteristics back into the lower-frequency range. This means that there is an inherent loss of information. The sampling process is reversible provided one has detailed knowledge of the relationship between the low-frequency and folded components so that they can be untangled from the sampled model. However, it is clear from the above argument that one has an inherent sensitivity to the assumptions that one makes about the folded components. The factors that contribute to the folded components include ¿ the sampling rate, ¿ the nature of the input between samples (i.e., is it generated by a firstorder hold or not, or is it continuous-time white noise or not), ¿ the nature of the sampling process (i.e., has an anti-aliasing filter been used and, if so, what are its frequency domain characteristics), ¿ the system relative degree (i.e., the high-frequency roll-off characteristics of the system beyond the base band), ¿ high-frequency poles and or zeros that lie outside the base band interval.

The usual recommendation made by clinicians to type 1 diabetics is that they should inject insulin when consuming a meal or, preferably, slightly earlier. However, in practice, insulin injection maybe delayed. In this paper we develop a fundamental limit on performance when insulin is delivered at some time other than the preferred time. The paper develops an optimal injection policy which minimizes the maximum blood glucose response whilst ensuring that the minimum response does not fall below a pre-specified level. The result provides a 'gold standard' against which other insulin injection policies can be compared. Implementation issues are also briefly described.

This paper describes the application of Model Predictive Control incorporating Stochastic Programming to the treatment of Type 1 diabetes. The use of stochastic programming in this context is believed to be important since it addresses one of the key difficulties associated with diabetes treatment namely the inherent uncertainty associated with future food and exercise patterns. Preliminary results are presented, based on real patients, illustrating the advantages of using stochastic model predictive control for this application.

Nonlinear System identification has a rich history spanning at least 5 decades. A very flexible approach to this problem depends upon the use of Volterra series expansions. Related work includes Hammerstein models, where a static nonlinearity is followed by a linear dynamical system, and Wiener models, where a static nonlinearity is inserted after a linear dynamical model. A problem with these methods is that they inherently depend upon series type expansions and hence it is difficult to know which terms should be included. In this paper we present a possible solution to this problem using recent results on rank-constrained optimization. Simulation results are included to illustrate the efficacy of the proposed strategy.

This paper proposes a new interpretation of Finite Set Model Predictive Control (FS-MPC) for inverters. This interpretation gives insights into the existing limitations of FS-MPC and suggests mitigation strategies. Consequently, significant improvements of FS-MPC performance are achieved, including elimination of steady state error, improved harmonic spectrum, reduced control delays, etc. The proposed new MPC interpretation also establishes relationships between various control schemes, including FS-MPC and PWM-MPC, FS-MPC and MPC with integral action, etc. Additionally, the paper proposes further embellishments to FS-MPC based on its more efficient and accurate implementation. The findings of the paper are supported by simulations and experimental results.

Model Predictive Control (MPC) has received significant attention from the power electronics research community. In this paper we apply MPC approach to designing optimal current regulators for voltage source inverters (VSI), in the sense of tracking references and rejecting disturbances of interest. In particular, we present a detailed derivation of optimal current regulators for tracking references and rejecting disturbances of constant and harmonic types. We compare these to the existing control structures, such as proportional integral (PI) and proportional resonant (PR) current regulators. We further propose a strategy to improve their performance based on matching their structure and parameters to the optimal controllers. The findings of the paper are supported by PSIM simulations and preliminary experimental results.

This paper discusses the question of saturating actuators in a multi-input multi-output (MIMO) control system design. It is shown that direct application of standard anti-windup schemes can lead to loss of closed-loop stability, as well as, the degradation of the directional properties of the system. Consequently, a novel anti-windup scheme is proposed, where the intended procedure ensures stability, accompanied with exact dynamic decoupling in the presence of actuator saturation. In essence, this is accomplished by ensuring that all of the control signals are saturated to the correct value, even-though if they do not saturate at all. Simulation results comparing the difference between the dynamic anti-windup methodology and a non anti-windup scheme are presented.

Model Predictive Control (MPC) has attracted significant attention from the power electronics community. In this body of work it is usual to utilize a linear load model. Also, to deal with computational complexity, it is common to deploy horizon 1 optimization. A major difficulty with these standard approaches is that modelling errors, e.g. those arising from non-linear effects such as magnetic saturation, lead to poor tracking performance. In this paper we show that perfect tracking is achievable provided that MPC controller is augmented by a suitable observer.

This paper develops a nonlinear insulin to carbohydrate rule for use by type 1-diabetes patients. The goal is to refine the commonly used Insulin to Carbohydrate Ratio (ICR) formula. The latter is a strictly linear rule relating carbohydrates consumed to insulin infusion. The new relationship presented in this paper is nonlinear and depends on the availability of a nonlinear dynamic model describing a patient's blood glucose response to food and insulin. Such a model can be obtained by use of nonlinear system identification tools applied to patient test data. The suggested procedure is of similar complexity to the existing standard ICR rule. Hence it has the potential to be of clinical importance especially in developing countries where sophisticated solutions such as an artificial pancreas are unlikely to be used due to excessive cost. Simulations are presented which show that there exists a significant difference between the suggested insulin provided by the rule developed here and that given by the standard ICR rule.

© 2014 IEEE.There is a very extensive literature on various aspects of the central Bias-Variance trade-off in linear system identification. In the 80's and 90's the focus was on bias characterization, model error models and Stochastic Embedding. Recently, there has been a new interest in Bayesian or kernel methods. This paper puts part of this literature into perspective by giving a modern interpretation of the Stochastic Embedding approach.

This paper presents a preliminary study of performance limitations that arise in the closed-loop control of blood glucose, using an autonomous artificial pancreas. It is shown that a major source of limitations is due to model uncertainty, specifically due to the combined effect of the insulin infusion system (IIS), the continuous glucose monitor (CGM) and the human glucose regulatory system. It is argued that the uncertainty associated with each of these elements compromises the achievable closed-loop bandwidth, and, in the presence of disturbances, e.g. meal intake and exercise, the closed-loop response will necessarily be poor. A proposition to overcome this problem is given based on feedforward action.

This paper addresses the control-rate versus quantizer-resolution trade-off in networked control. The case presented considers the situation where the bit rate between controller and plant is constrained to a fixed number of bits per unit of time and an underlying fixed fast sampling rate is deployed to take measurements. However, a variable control update rate can be used between the controller and the plant. Inspired by the practical problem of inner loop power control in WCDMA, we assume the plant is an integrator. This assumption covers more general plants when fast sampling is used. Also, a restricted architecture in which linear filters are used for the encoder/decoder is considered and a quantizer with linear feedback is deployed. These choices give maximal insights into the underlying problem. It is shown that, in this case, it is best to use one bit per sample, in which case, the control update frequency is equal to the bit rate.

This paper studies the problem of channel estimation for the downlink in broadband LTE communications. A Kalman Filter is used to combine channel soundings made with pilot tones that are deployed at different times and in different frequency bands on a pre-selected pattern. This Kalman filter is presented in a form directly suitable for implementation as a periodically time varying filter, in the case where periodic pilot tone data is available. Simulation results illustrate the efficacy of this approach.

We study a state estimation architecture for sensor networks, where several sensors transmit quantized innovations to a central estimator. Transmission is via a wireless channel, which is prone to fading leading to random packet loss. State estimation is carried out at the gateway via a time-varying Kalman filter which accounts for packet loss and quantization effects. To form the innovations at the sensors, the estimator transmits information regarding its current state estimate to the sensors. This information could be dedicated to each sensor or broadcast to all sensors. In addition, the gateway also decides upon power levels and quantization step-sizes to be used by each sensor node. Here, we adopt elements of predictive control to trade off estimation performance versus energy use.

This paper studies LTI one-degree-of-freedom networked control architectures built around LTI SISO plants. The feedback path comprises an unreliable channel that drops data in an i.i.d. fashion. We extend known results to show that there exists a second order moments equivalence between the situation examined and an auxiliary feedback loop where the unreliable channel has been replaced by an additive i.i.d. noise channel that has an equality signal-to-noise ratio constraint. We then use these insights to design optimal controllers. We also establish a necessary and sufficient condition on the successful transmission probability that allows for the design of a controller that guarantees MSS in the considered setup. Our proposal does not require packet arrival acknowledgements. Thus, our results give necessary and sufficient conditions for MSS in an output-feedback control architecture with UDP-like protocols.

Advanced design aims to achieve the best possible performance subject to operational constraints. Such questions can be formulated as optimization problems. The advantages of using an explicit optimization formulation include clear articulation of performance trade-offs and the provision of a clear basis for quantitative comparison of different strategies. In this paper we will briefly review aspects of modern optimization. We will also show how these tools can be deployed in scheduling, estimation and control. We will illustrate the ideas by examples drawn from the mining, mineral and metal processing field. Copyright © 2007 IFAC.

We present a receding horizon control algorithm for compensation of backlash at the input of a stable linear system under control rate constraints. The problem is first posed as a receding horizon optimal control problem by modelling backlash as a piecewise affine system with a state space partition consisting of three regions. This optimal control problem involves solving, at each step, 3N quadratic programmes, where N is the optimisation horizon. As an alternative to solving the quadratic programmes, we propose a strategy based on a recently devised suboptimal receding horizon control algorithm which utilises a singular value decomposition of the Hessian of the quadratic programme. This alternative strategy leads, at the cost of some performance degradation, to much smaller computational load since a feasible rather than optimal solution has to be obtained at each step. Copyright © 2005 IFAC.

A well known difficulty in system identification over a large bandwidth is the ill-conditioning of the normal matrix. This typically manifests itself as poor or erroneous estimates. Several methods have been proposed in the literature for addressing this issue. However, none appear to give an entirely satisfactory solution. Here we present a novel technique, utilising particular basis functions, aimed specifically at improving the numerical properties of the least squares normal matrix in transfer function estimation over a wide bandwidth. We show that, under some mild assumptions, the achieved condition number is actually independent of the frequency range. Several examples are presented showing the superior performance of the proposed method when applied to wide-band estimation problems.

We address the problem of finite-horizon optimal control of uncertain discrete-time SISO linear systems with input constraints. The uncertainty for the problem analyzed is related to incomplete state information (output feedback) and stochastic disturbances. We analyze the optimal solutions for short optimization horizons when the disturbances and the initial state are Gaussian random vectors. We also consider two suboptimal strategies, one of which is known as Certainty Equivalent Control.

In this paper, we revisit the factors that limit the achievable performance of rudder-based roll stabilizers for ships. We quantify the best performance in terms of roll reduction, and roll reduction vs. heading interference for a given sea state and sailing condition. We do this in the context of limiting optimal control incorporating input constraints. The performance indices obtained provide a benchmark against which other control strategies can be compared, and also provide a criterion to assess applicability of the this type of stabilizer to a particular ship.

In high sea states, the effectiveness of fin stabilizers can severely deteriorate due to non-linear effects arising from unsteady hydrodynamic characteristics of the fins; namely dynamic stall. In this paper, we review some concepts related to this phenomenon, and suggest the application of a constrained control strategy to prevent the development of these non-linear effects for fin stabilizers.

The solution to constrained linear model predictive control (MPC) problems can be pre-computed off-line in an explicit form as a piecewise affine (PWA) state feedback law defined on polyhedral regions of the state space. Even though real-time optimization is avoided, implementation of the PWA state-feedback law may still require a significant amount of computation due to the problem of determining which polyhedral region the state lies in. In this paper, a neural network approach to this problem is investigated.

This paper analyzes the conditions for local convergence of Maximun Likelihood estimation for ARMAX models. We do this by examining the region in which the steepest descent direction leads to a reduction in the Euclidean norm of the parameter error. Inter-alia this gives new insights into the question of existence of local maxima of the likelihood function for various commonly used model structures.

This paper considers the singular structure for sampled data LQ problems of continuous linear systems, and the asymptotic behaviour when the sampling rate is increased. We show that there is a natural convergence between the finite set of singular values of the discrete time problem, and the (infinite) countable set in continuous time.

Model predictive control is a well established strategy for dealing with saturating actuator problems. The essential feature of the method is a receding horizon quadratic optimal control problem which is solved subject to input constraints. On the other hand, anti-windup methods are ad-hoc procedures which achieve input saturation in an instantaneous fashion. Both methods are known to perform well in practice and each has its strong advocates. In this paper, we explore connections between the methods. In particular, we show that there are cases in which anti-windup schemes are identical to MPC schemes. In other cases, we show that anti-windup has performance which is close to that of MPC strategies.

To reduce stored capacitor energy in modular multilevel converters (MMCs), previous papers propose injection of harmonics in the arm circulating currents and zero-sequence voltages. While this approach is effective, previously utilized combinations of injected harmonics significantly increase converter current ratings. In contrast, the optimization approach developed in this article identifies harmonics that simultaneously achieve capacitor energy optimization and reduced overrating of converter currents. In addition, the proposed approach eliminates the possibility of convergence to local minima, which is a significant improvement on previously proposed nonconvex optimization approaches. Experimental verification of the proposed approach is performed on a low-voltage three-phase modular multilevel converter.

Standalone microgrids are used in various industrial operations to supply critical loads. The authors have proposed a stable and robust decentralized ac microgrid architecture that provides an automatic and equitable load sharing between distributed generators, without the use of droop control. To deal with faults or real-time changes in the distributed generation levels, a flexible reconfiguration of the microgrid can be advantageous. To achieve that, this article proposes the use of power line communication (PLC). Two PLC strategies are explored in this article: the first is based on using a single and constant PLC frequency, and the second is based on using a variable PLC frequency within a given range. This article details PLC injection, detection, and suppression methods, as well as the overall communication logic. It demonstrates the ability of the microgrid to adapt to changes while minimizing the amount of battery support. Findings of this article are supported by simulations and validated by experiments.

Matrix converter is an attractive choice for motor drives with weight/size restrictions and high efficiency requirements, such as electric haulage vehicles used in mining. However, the use of matrix converters requires addressing a potential issue with input resonance, which can be exacerbated when using long feeders or trailing cables. This article addresses this problem by proposing a novel solution based on principles of modern control but shaped as a classical feedback architecture. Along with high quality reference tracking, the proposed solution utilizes the frequency changing properties of matrix converters to attenuate the input filter resonance by spectral shaping of the output current. This article shows that the solutions achieves its goals. Design and implementation details are discussed. This article results are validated by simulations and experiments.

Open-loop poles on, or near, the imaginary axis in the complex plane are a common occurrence in feedback systems. They can occur as part of the model in systems containing lightly damped resonances or they can be deliberately introduced so as to ensure perfect tracking for sinusoidal or periodic signals. This paper establishes both time and frequency domain implications of such poles on system performance. Dual results for zeros on, or near, the imaginary axis are also discussed. Many practitioners are, at least heuristically, aware of the broad implications of these results but, to the best of the authors' knowledge, they have not previously been explicitly articulated nor quantified.

The estimation of a vehicle's position and velocity based on measurements of pseudoranges and range-rates is a nonlinear filtering problem. We utilize a linearized Kalman filter with global convergence that avoids divergence because the local model linearization is not based on the linearized Kalman filter's own estimates. Instead, the linearization is made about the estimates of another suboptimal Kalman filter that is based on a globally valid quasi-linear time-varying measurement model.

Model predictive control (MPC) has been widely advocated as a design strategy for many aspects of industrial electronics. The methodology has been strongly promoted by some researchers but has also attracted criticism from others. In this context, the purpose of this paper is twofold. First, we show that many existing and popular control strategies, including finite set MPC and linear controllers [proportional integral, proportional resonant (PR)], can be viewed as special cases of MPC. Second, we show that the predictive control framework allows one to embellish these classical control architectures with novel features and to design new and advanced control architectures to address various challenges posed by power electronics applications. The findings of the paper are supported by a practical example of designing of a novel form of PR controller with superior tracking performance and delay compensation, confirmed via simulation and experiments.

© 2014 Taylor and Francis. Inverters play a central role in modern society including renewable energy integration and motor drives. Due to the inherent switched nature of the inverter waveforms harmonic distortion is an issue. Additionally, the switching patterns are perturbed by unavoidable switching delays. Amongst those, nonlinear and load-dependent switching delays (known as inverter dead-time delays) are the most difficult to compensate. In this paper, we propose a new approach to delay compensation and harmonic suppression in inverter voltage. The proposed approach is based on variable prediction horizon nonlinear model predictive control.

In networked control, there is often an incentive to communicate only what is absolutely necessary to achieve the desired performance goals. This is true to both the downlink (between a control base station and actuators) and the uplink (between the sensors and base station). Here, we present a strategy aimed at this problem based on a singular value decomposition of the Hessian of the quadratic performance index generally considered in Model Predictive Control. The singular vectors are employed to generate an orthonormal basis function expansion of the unconstrained solution to the finite horizon optimal control problem. These are preloaded into each actuator and each sensor. On the downlink, the actuators are informed, in real-time, about which basis functions they should use. On the uplink, after a 'burn in period', the sensors need only communicate when their response departs from that pre-calculated for the given basis functions. We show that this strategy facilitates communication in both the downlink and uplink in a cost-effective fashion. We also show that the strategy can be modified so that input constraints are satisfied. We illustrate the proposed results by applying them to a simulation of the cross direction control of a paper machine. Potential extensions and other applications of the proposed strategy are also discussed.

In this paper we establish a relationship between the complementary sensitivity function of output feedback control and two other sensitivity functions, namely a (scaled) complementary sensitivity function associated with feedback control based on an auxiliary variable, and a complementary sensitivity function associated with an observer. Implications of the results in related problem areas, e.g. robust model predictive control, are briefly discussed.

This paper presents an identification scheme for sparse FIR systems with quantised data. We consider a general quantisation scheme, which includes the commonly deployed static quantiser as a special case. To tackle the sparsity issue, we utilise a Bayesian approach, where an l1 a priori distribution for the parameters is used as a mechanism to promote sparsity. The general framework used to solve the problem is maximum likelihood (ML). The ML problem is solved by using a generalised expectation maximisation algorithm. © 2013 Taylor & Francis.

This paper addresses the following problem in networked control: "If a control law is implemented over a channel that supports a certain fixed bit rate what is the best choice for the control update rate and, consequently, the number of bits carried in each sample?" A restricted architecture in which linear filters are used for the encoder/decoder is considered and a quantizer with linear feedback is deployed. Subject to these restrictions, a procedure for designing the controller and associated filters is presented. These filters are then deployed to choose the best number of bits per control update. It is shown, subject to the above restrictions, that it is generally best to use one bit per sample, in which case, the control update rate is equal to the bit rate. Our analysis has two points of departure from contemporary literature in this area. Firstly, we focus on bits per unit time, as opposed to bits per sample. Secondly we use a fixed number of bits in every time period as opposed to an average bit rate. © 2014 Elsevier Ltd. All rights reserved.

This paper addresses the following problem in networked control: "If a control law is implemented over a channel that supports a certain fixed bit rate what is the best choice for the control update rate and, consequently, the number of bits carried in each sample?" A restricted architecture in which linear filters are used for the encoder/decoder is considered and a quantizer with linear feedback is deployed. Subject to these restrictions, a procedure for designing the controller and associated filters is presented. These filters are then deployed to choose the best number of bits per control update. It is shown, subject to the above restrictions, that it is generally best to use one bit per sample, in which case, the control update rate is equal to the bit rate. Our analysis has two points of departure from contemporary literature in this area. Firstly, we focus on bits per unit time, as opposed to bits per sample. Secondly we use a fixed number of bits in every time period as opposed to an average bit rate. © 2014.

In spite of its easy implementation, ability to handle constraints and nonlinearities, etc.; model predictive control (MPC) does have drawbacks including tuning difficulties. In this paper, we propose a refinement to the basic MPC strategy by incorporating a tuning parameter such that one can move smoothly from an existing controller to a new MPC strategy. Each change of this tuning parameter leads to a new stabilising control law, therefore, allowing one to gradually move from an existing control law to a new and better one. For the infinite horizon case without constraints and for the general case with state and input constraints, stability results are established. We also examine the practical applicability of the proposed approach by employing it in the nominal prediction model of the tube-based output feedback robust MPC method. The merits of the proposed method are illustrated by examples. © 2013 Elsevier Ltd. All rights reserved.

In this paper, we address channel-impulse response (CIR) estimation in multicarrier systems with phase distortion, namely, phase noise (PHN) and carrier frequency offset (CFO). The estimation problem also considers the joint estimation of the channel noise variance, CFO, and PHN bandwidth. We develop a general state-space model for multicarrier systems, separating the complex signals into their real and imaginary parts. This provides a valid framework for any modulation scheme (proper or improper). We use the expectation-maximization (EM) algorithm to solve the maximum-likelihood (ML) estimation problem. Our approach exploits the linear and Gaussian structure associated with the transmitted signal. Due to the nonlinear nature of the PHN, sequential Monte Carlo (MC) techniques are considered. Our analysis includes general expressions under different training scenarios.We show, via simulation, the impact of PHN bandwidth estimation on overall parameter estimation, and we study the impact of different training levels. In addition, we consider the accuracy of the parameter estimates, providing expressions for the Fisher information matrix (FIM) and focusing on the estimation accuracy of the PHN bandwidth. © 2012 IEEE.

In this paper, we propose a novel approach to nonlinear filtering utilizing on-line quantization. We develop performance bounds for the algorithm. We also present an example which illustrates the performance of the method. © 2013 Copyright Taylor and Francis Group, LLC.

Modern signal processing and control algorithms are invariably implemented digitally, yet most real-world systems evolve in continuous time. Hence, the interaction between sampling and the behavior of continuous-time systems is an important ingredient in all real-world signals and systems problems. © 2012 IEEE.

Induction Heating Furnaces are used extensively in industry. The basic principle is that induced eddy currents are used to heat a ferromagnetic material as it passes through a series of coils. Because of the importance of such systems, there has been on-going interest in their design and operation. Past work includes model development from physical principles and optimal design of operational practices. However, previous work has invariably been based on open-loop strategies. Our work is aimed at the design of a closed-loop control strategy incorporating feedback from the available measurements. This paper reports initial work including model development and calibration together with preliminary control system design. Proposed future work includes full scale industrial implementation.© 2013 Elsevier Ltd. All rights reserved.