Decoupling and determination of the inverse for time-varying multivariable systems

Chang transformation was introduced for decoupling the slow and fast dynamics of a singularly perturbed Linear Time Invariant (LTI) system, and it was subsequently extended to Linear Time-Varying (LTV) systems under the slowly-varying assumption by way of frozen-time eigenvalues. This paper extends Chang transformation from slowly-varying LTV systems to LTV systems, when the singularly perturbed system has a semi-proper coefficient matrix A <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">22</sub> (t) and LTV system ẇ(t) = A <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">22</sub> (t)w(t) is exponentially stable. Instead of using frozen-time eigenvalues based on slowly time-varying conditions, PD-eigenvalues for LTV systems are employed to characterize the exponential stability, thereby circumventing the slowly-varying constraint. Our results provide a larger bound on epsilon, which is a gauge on the validity of the Chang transformation, in some situations than using previous techniques. We have also recast the original Chang transformation so as to provide more insight into the decoupled subsystems. The insight will be useful in subsequent investigation on estimate of the Singular Perturbation Margin (SPM) for LTV systems. An equivalence relationship has recently been established for LTI system between the SPM and the Phase Margin (PM). The new results in this paper will facilitate the development of a PM-type stability margin metric for LTV systems, and for nonlinear, time-varying systems in the future.

A new designing method of multivariable linear time-varying tracking systems with stochastic disturbances for polynomial command inputs is proposed. It has been shown that it is possible to choose the matrices of gain elements placed in feedbacks and parallel paths of the closed-loop system so that the output y(t) tracks the polynomial command input v(t) with (E expectation operator).

A method of generating matrices N(t) for transforming a linear multivariable time-varying system to the companion form has been given. In some cases, the transformation matrix could be made time invariant.

An identification approach for linear and nonlinear time-variant structural systems via harmonic wavelets

ABSTRACTIn this paper, we describe a constructive test to decide whether a given linear time-varying (LTV) differential system admits a stabilising compensator for the control tasks of tracking, disturbance rejection or model matching and construct and parametrise all of them if at least one exists. In analogy to the linear time-invariant (LTI) case, the ring of stable rational functions, noncommutative in the LTV situation, and the Kučera–Youla parametrisation play prominent parts in the theory. We transfer Blumthaler's thesis from the LTI to the LTV case and sharpen, complete and simplify the corresponding results in the book ‘Linear Time-Varying Systems’ by Bourlès and Marinescu.

An algorithm, based on numerical techniques, for calculating the open-loop optimal control for time-varying linear multivariable systems is considered. An approximation to the open-loop control is developed to provide a suboptimal closed-loop controller which is shown to be close to controllers computed using the maximum principle or dynamic programming procedures. Application to a nonlinear level control system shows that the measured response is close to the predicted optimal response for disturbances of up to 25% from the operating point.

An algorithm, based on the fixed-point contraction-mapping principle of functional analysis, for calculating the optimal controls for time-varying linear multivariable systems is considered. By an appropriate transformation of the operators involved, it is shown that the convergence rate may be maximised and that the resulting optimal algorithm will always converge. Comparison of the optimal algorithm with the steepest-descent and conjugate-gradient schemes shows the optimal algorithms to have significantly superior convergence rates, as well as requiring appreciably smaller computer storage.

In this paper, adaptive optimal control is proposed for linear discrete time-varying (LDTV) systems subject to unknown system dynamics. The idea of the method is a direct application of the Q-learning adaptive dynamic programming for time-varying systems. In order to derive the optimal control policy, an actor-critic structure is constructed and the time-varying least square method is adopted for parameter adaptation. The derived control policy robustly stabilizes the time-varying system and guarantees an optimal control performance. As no particular system information is required throughout the process, the proposed method provides a feasible solution to a large variety of applications. The validity of the proposed method is verified through simulation studies.

In this paper a new systematic switching control approach to adaptive stabilization of linear time-varying (LTV) discrete-time systems is presented.A feature of the localization-based method is its high model falsification capability, which in the case of LTI systems is manifested as the rapid convergence of the switching controller.We believe that the proposed method may help pave the way for design of practical adaptive switching controllers applicable to a wide range of linear time-invariant and timevarying systems.

Periodic Motions for Estimation of the Attraction Domain in the Wheeled Robot Stabilization Problem

Time-varying pascal systems

A physical parameter identification method for the linear time-varying system is presented in this paper. For an arbitrarily time-varying system, a series of Hankel matrices are established directly from the combined measurements of displacement, velocity, and acceleration together with the input excitations from a single experiment with the system. The time-varying equivalent state-space system matrices of the structure at each time instance are then identified by singular value decomposition using the free or random responses. A time-varying transformation matrix for transforming the identified equivalent system matrices into the physical system matrices is developed to determine the mass, stiffness, and damping matrices. Details of the proposed approach are investigated with a two degrees-of-freedom spring-mass-damper system with three kinds of time-varying parameters. Numerical results show that the proposed method is accurate and effective in identifying time-varying physical parameters.

The paper deals with the construction of generalized frozen-time (FT) models-for linear time-varying continuous-time systems. The generalized FT models are higher-order models generated from a given n-dimensional state model by applying a stable differential operator to the state equation. This results in FT models whose coefficients depend on the derivatives of the time-varying system coefficient matrices. The accuracy of the higher-order FT models over local time intervals is investigated in terms of truncated Taylor series. In the last part of the paper, the application of the higher-order FT models to state feedback control is considered.

In the case of multiple nonstationary independent source signals and linear instantaneous time-varying mixing systems, it is difficult to adaptively separate the multiple source signals. Therefore, the adaptive blind source separation (BSS) problem is firstly formally expressed and compared with tradition BSS problem. Then, we propose an adaptive blind identification and separation method based on the variable learning rate equivariant adaptive source separation via independence (EASI) algorithm. Furthermore, we analyze the scope and conditions of variable-learning rate EASI algorithm. The adaptive BSS simulation results also show that the variable learning rate EASI algorithm provides better separation effect than the fixed learning rate EASI and recursive least-squares algorithms.

For linear time-varying single-input single-output systems, the paper treats the problem of designing an output feedback controller resulting in an (internally) stable closed-loop system that tracks a given reference signal. The design is accomplished by augmenting the given system to include an internal model of the reference signal. The internal model reduces to the usual form when the given plant is time invariant. An explicit construction for the internal model is given in the case of a step reference.