Stable Inversion of Discrete-Time Linear Periodically Time-Varying Systems via Cyclic Reformulation
Hiroshi Okajima · 2026
Read Stable Inversion of Discrete-Time Linear Periodically Time-Varying Systems via Cyclic Reformulation by Hiroshi Okajima (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-23147.txt.
Introduction
Okajima derives a stable inverse for discrete-time linear systems that vary in a periodic way. Such inverses serve feedforward and iterative learning control. His route uses cyclic reformulation, which turns a periodic system into a matching time-invariant one. A structure-preservation result anchors the proof. The invariant inverse keeps the cyclic block form, so periodic blocks can be read off one by one. The inverse is real-valued and causal or delayed by set steps. Stability matches the plant's periodic invariant zeros lying in the unit disk.
Catalog record
| CREATOR | Hiroshi Okajima |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 56480 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2603.23147
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Stable Inversion of Discrete-Time Linear Periodically Time-Varying Systems via Cyclic Reformulation Thanks: Corresponding author: H. Okajima. Email: [email protected] Hiroshi Okajima Abstract Stable inverse systems for periodically time-varying plants are essential for feedforward control and iterative learning control of multirate and periodic systems, yet existing approaches either require complex-valued Floquet factors and noncausal processing or operate on a block time scale via lifting. This paper proposes a systematic method for constructing stable inverse systems for discrete-time linear periodically time-varying (LPTV) systems that avoids these limitations. The proposed approach proceeds in three steps: (i) cyclic reformulation transforms the LPTV system into an equivalent LTI representation; (ii) the inverse of the resulting LTI system is constructed using standard LTI inversion theory; and (iii) the periodically time-varying inverse matrices are recovered from the block structure of the cycled inverse through parameter extraction. For the fundamental case of relative degree zero, where the output depends directly on the current input, the inverse system is obtained as an explicit closed-form time-varying matrix expression. For systems with periodic relative degree r ≥ 1 r\geq 1 , the r r -step-delayed inverse is similarly obtained in explicit closed form via the periodic Markov parameters. The stability of the resulting inverse system is characterized by the transmission zeros of the cycled plant, generalizing the minimum phase condition from the LTI case. Numerical examples for both relative degree zero and higher relative degree systems confirm the validity of the stability conditions and demonstrate the effectiveness of the proposed framework, including exact input reconstruction via causal real-valued inverse systems. Keywords: Inverse systems; Periodically time-varying systems; Cyclic reformulation; Stable inversion; Multirate systems. 1 Introduction The inverse of a dynamical system plays a fundamental role in a wide range of control and signal processing applications: feedforward control and output tracking [ 1 , 5 ] , iterative learning control (ILC) [ 6 ] , and invertibility analysis of nonlinear systems [ 2 ] . For linear time-invariant (LTI) systems, stable inversion theory is well developed: if all transmission zeros lie strictly inside the unit disk (the minimum phase condition), a causal stable inverse exists and can be constructed explicitly in state-space form [ 3 , 4 ] . When nonminimum phase zeros are present, a bounded noncausal inverse can be obtained via exponential dichotomy [ 11 , 9 ] , at the cost of requiring future output information (preview). Many practical systems, however, exhibit periodically time-varying dynamics that cannot be captured by an LTI model. Linear periodically time-varying (LPTV) systems arise naturally in multirate sampled-data control [ 22 ] , position-dependent and non-equidistantly sampled systems [ 8 , 9 ] , and high-precision positioning stages [ 10 ] . Because the system matrices vary periodically, the transfer function framework underlying LTI inversion theory is no longer available, making the construction of stable inverse systems considerably more involved. Existing work on LPTV inversion falls into three main categories; a comprehensive comparison of inversion-based approaches for feedforward and ILC, including extensions to time-varying systems, is given in [ 12 ] . The first is the stable inversion approach of van Zundert and Oomen [ 8 , 9 ] , developed for discrete-time LPTV systems. This approach adapts the exponential dichotomy framework of Devasia and Paden [ 11 ] , originally developed for continuous-time nonlinear systems, to the discrete-time periodic setting via a Floquet-type factorization of the monodromy matrix. The resulting inverse operator is defined on the doubly-infinite time axis and is exact in the infinite-horizon sense. However, it requires complex-valued Floquet factors and noncausal processing, and does not yield a closed-form state-space realization suitable for direct causal implementation. The second category consists of lifting-based approaches for discrete-time LPTV systems. Zhu et al. [ 13 ] apply the standard lifting technique to transform a periodic system into an equivalent LTI system, then construct the stable inverse of the lifted system by combining stable inversion with optimal state transitions. More recently, Zhu et al. [ 14 ] propose a double lifting approach that handles nonminimum phase multivariate LPTV systems by formulating the inversion as a finite-dimensional matrix equation and applying the Moore–Penrose pseudoinverse, guaranteeing L_{2} -bounded inputs for finite-time tracking. However, since both approaches are based on lifting, the resulting inverse operates on the N -step block time scale and produces a finite-time input sequence rather than a step-by-step periodic state-space realization in the original time domain. The third category consists of signal-processing-oriented methods for discrete-time LPTV filters [ 16 , 15 , 17 ] , which focus on FIR/IIR filter inversion using polyphase or modulator representations, or on stable/unstable decompositions of LPTV digital filters. These methods are not formulated in a state-space control framework and typically assume the direct feedthrough term is nonzero (relative degree zero). Invertibility of linear time-varying discrete-time systems was studied in a foundational paper by Kono [ 7 ] , which provides existence conditions for left and right inverses of general LTV systems in terms of rank conditions on structured matrices. This provides the theoretical underpinning for the inversion problem but does not give a constructive algorithm specialized to the LPTV case. A closely related line of work concerns the model matching problem for periodic systems. Colaneri and Kučera [ 18 ] showed that, via cyclic reformulation, a periodic state-feedback law can be designed so that the closed-loop input–output map matches a prescribed target transfer function. The zero-matching conditions in their framework are expressed through the cyclic transfer function and are thus closely related to the periodic transmission zero conditions that govern the stability of the inverse system in the present work. Tange and Tsumura [ 19 ] extended this perspective by formulating periodically weighted model matching problems using dual lifted forms for causal LPTV controllers, and showed that LPTV controllers can outperform LTI ones in rejecting periodic disturbances. The inversion problem studied here can be viewed as a special case of model matching in which the target is the identity map; the key difference is that the present work derives an explicit periodic state-space realization of the inverse rather than a state-feedback gain. The present paper proposes a fundamentally different approach based on cyclic reformulation [ 20 , 21 ] . Cyclic reformulation transforms an N -periodic LPTV system into an equivalent LTI system by a state, input, and output coordinate transformation that exploits the periodic structure. The key advantage over lifting [ 22 ] is that cyclic reformulation operates at the original sampling rate (rather than N -step block time) and preserves the step-by-step input-output relationship, while still enabling the direct application of LTI tools. This capability has been exploited to extend LTI design methods to the periodic setting without conservatism, including LMI-based controller design for multirate systems [ 23 ] , multirate Kalman filter design with semidefinite noise covariances [ 24 ] , and system identification of periodic plants [ 25 ] . The proposed inversion procedure consists of three steps: (i) cyclic reformulation transforms the LPTV plant into a cycled LTI system \check{P} ; (ii) the inverse \check{P}^{-1} is constructed using standard LTI inversion theory; and (iii) the periodically time-varying inverse matrices are extracted from the block structure of \check{P}^{-1} through parameter extraction. This procedure yields explicit closed-form