Search

This paper presents a robust tracking model predictive control (MPC) strategy for offset-free regulation of the output of input-constrained uncertain systems to a non-zero reference signal. To this end, a feedback control law with integral action is proposed, and a feasible and invariant set is explicitly derived on the basis of the feedback law. Finally, a robust one-step-ahead tracking MPC is devised on the basis of the feasible and invariant set in order to improve the control performance. Simulation results show that the proposed tracking MPC successfully achieves robust tracking in terms of both control performance and the size of the feasible set. [ABSTRACT FROM PUBLISHER]

In this article, the static output feedback problem for linear time-invariant systems is considered. For arbitrary assignability of the roots of the characteristic polynomial by static output feedback, a new necessary and sufficient condition is derived. Although, the proof is based on simple analysis, the known sufficient conditions (derived by techniques of algebraic geometry) are directly covered. Furthermore, an algorithm for the calculation of feedback matrices assigning a desired set of eigenvalues is proposed. This algorithm does not require the desired eigenvalues to be distinct and it explicitly exploits the available degrees of freedom for reducing the feedback gain. The presented approach is illustrated on computational examples. [ABSTRACT FROM PUBLISHER]

Model reference adaptive control is a major design method for controlling plants with uncertain parameters. The primary objective of this paper is to develop a new design approach for a differentiator-free model reference adaptive control of a single-input single-output linear time-invariant plant. The proposed method, called the “Identifier-tracking model reference adaptive control”, uses a stacked identifier structure that is new to the field of adaptive control. The goal is to make the output of the plant asymptotically track the output of the first identifier, and then driving the output of the first identifier to track that of the second identifier, and so forth, up to the qth identifier where q is the relative degree of the plant. Lastly, the output of the qth identifier is forced to converge to that of the reference model. Simulation results show the superiority of the proposed method over the traditional model reference adaptive control with augmented error in terms of the transient response. Since the resulting control systems are non-linear and time-varying, the stability analysis of the overall system plays a central role in developing the theory. [ABSTRACT FROM AUTHOR]

This paper develops a methodology for actuator fault diagnostics and quantitative estimation of fault signals in a class of non-linear systems. The class of non-linear systems considered is one in which the non-linearity is an incremental quadratic type of non-linearity that includes Lipschitz, positive real and sector non-linear functions. The methodology provides for both state estimation and fault vector estimation. An LMI procedure can be utilized for explicit computation of the observer gains. The procedure developed can also be specialized to linear time invariant systems. Compared to previous results for LTI systems, the procedure developed herein does not require the LTI system to be minimum phase. The use of the developed methodology is demonstrated through two illustrative examples of real world physical applications. [ABSTRACT FROM AUTHOR]

Robust non-fragile feedback stabilization is studied for causal linear time-invariant (LTI) input–output systems defined in an L 2 () signal space setup, or equivalently in an L 2 (j) setup in the frequency domain. Two-operator descriptions for the plant and the controller are used with bounded operators defined on the full space L 2 (). Basic properties of causal LTI operators on weighted L 2 spaces, L 2 (,   w ), which are relevant to stabilization theory, are also established for different choices of the weighting function w . Weighted L 2 spaces allow the inclusion of persistent signals such as steps and thus provide the potential of important generalizations of the robust stabilization theory on L 2 (). In particular, it is shown that there is a large class of weights w for which the L 2 (,   w ) setup leads to H ∞ optimization. This implies that robust H ∞ design can be applied to a large class of weighted L 2 spaces on . Problems of closability of unstable causal LTI convolution operators are discussed. This strengthens the importance of the two-operator model Ay   =   Bu as the basic plant input–output model on the full time axis. It is discussed how to obtain such models with A and B bounded causal coprime operators defined on the full signal space. [ABSTRACT FROM AUTHOR]

In this paper, we address some robustness issues for a class of switched linear systems with non-linear perturbations. We illustrate through an example that the well-known state-space-partition-based switching law may lead to chattering in the event that the systems undergo perturbations. To overcome this, we propose a state-feedback switching law with a non-zero level set. We prove that this switching law can prevent the system from chattering. In addition, this switching law makes the perturbed systems bounded with bounded perturbations, and practically convergent with convergent perturbations. We also propose an observer-based switching law to steer the switched system stable with nice robustness performances. [ABSTRACT FROM AUTHOR]

The paper presents a design method for optimizing the loop-gain of unstable-non-minimum phase feedback systems. Because obvious gain-bandwidth limitations exist in such systems, the proposed best loop gain optimization goal is maximization of the stability gain and phase margins, which are inherently low in such feedback systems. The design procedure is based on coprime factorization of the plant and controller, and the Bezout identity. An optimization process varies iteratively the observer-based controller parameters until maximization of the gain and phase margins is achieved. The same design technique also allows one to solve the mixed sensitivity H ∞ standard control problem for linear time invariant feedback systems. [ABSTRACT FROM AUTHOR]

Three new qualitative control concepts are introduced and developed in the framework of general MIMO linear, time-invariant, continuous time systems to be controlled: natural trackability, elementwise and stablewise tracking, and natural tracking control. The study of these concepts opens new problems in control science, which are solved in the form of their necessary and sufficient conditions. Different physical natures and/or purposes of the system output variables, as well as their mutual independence, may require that each output be assigned a specific settling time and/or a reachability time. Such times compose corresponding vector times. The natural tracking control ensures elementwise output tracking which obeys a prespecified vector settling time and/or vector reachability time. A controller should use information only about the system output error and the realized control itself in order for the control to be natural tracking control. But the controller does not use any information about the system internal dynamics description, system state and external disturbances. The necessary and sufficient conditions are proved for the stability of systems that exhibit tracking. They permit possible independent and separate synthesis of a stabilizing control and of a tracking control. A natural tracking control synthesis and natural tracking controller design for a fourth-order MIMO system and simulation results illustrate and verify the new control theory avenues. [ABSTRACT FROM AUTHOR]