Power-based control of output oscillations with online estimation of biased harmonics Authors accepted manuscript

Michael Ruderman, Denis Efimov · 2025

Read Power-based control of output oscillations with online estimation of biased harmonics Authors accepted manuscript by Michael Ruderman, Denis Efimov (2025) 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-2504-16445.txt.

Introduction

Michael Ruderman and Denis Efimov improve a lean control method for shaking outputs in this two thousand twenty-five paper. Some systems settle into slow, stubborn oscillations. A power-based scheme damps them while sending very few control signals: it acts only twice per swing, at the peaks. That thrift saves communication effort in the loop. The scheme needs the swing's frequency, size, and offset at every moment.

The authors add a fast estimator that learns those values online. They also sharpen the formula for the control impulse's weight. Tests ran on a fifth-order actuator with a load hanging free under gravity, with sensor noise present. The work serves control engineers with tight bandwidth. Accepted manuscript notes mark this as the journal-bound version.

Catalog record

CREATORMichael Ruderman, Denis Efimov
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE24 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2504-16445

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Power-based control of output oscillations with online estimation of biased harmonics Authors accepted manuscript Michael Ruderman, Denis Efimov Abstract The recently introduced discrete power-based control ( Ruderman 2024b ) reduces largely the communication efforts in the control loop when compensating for the marginally damped or even slowly diverging output oscillations. The control commutates twice per oscillations period (at the amplitude peaks) and uses the measured harmonic output only. The power-based control scheme requires the knowledge of the instantaneous frequency, amplitude, and bias parameters of the harmonic signal. This paper extends the power-based control by the finite-time estimation of the biased harmonics ( Ahmed et al. 2022 ). Also an improved analytic calculation of the impulse weighting factor is provided. The power-based oscillations control with online estimation of the harmonic parameters is evaluated experimentally on the fifth-order actuator system with a free hanging load under gravity and measurement noise. keywords Power based control \sep oscillations compensation \sep frequency estimation \sep biased harmonics \sep finite-time parameters estimation \sep discrete feedback control , 1 Introduction The control, or often just a mitigation, of the harmonic oscillations and/or vibrations in the output state is an essential issue for various mechanical systems and application scenarios. Examples can be found in e.g. active suspensions ( Landau et al. 2005 ), off-shore rotary drilling systems ( Wang et al. 2020a ), actuator drives with elastic connections ( Ruderman et al. 2012 ), two-rotor vibrational systems ( Fradkov et al. 2020 ), to mention here a few. The associated control problems are often exacerbated by the fact that sensor and actuator elements are not co-located, like in case of e.g. drill string systems ( Aarsnes and van de Wouw 2018 ) or large flexible structures ( Gibson et al. 2011 ). The non-collocated oscillatory systems can suffer under a limited access to the internal dynamic states and a poor observability, that makes an observer design and full-state feedback control hardly achievable, as demonstrated in Ruderman 2024a . If a non-collocated sensor/actuator configuration is in place, a reduced communication rate in the closed control loop can also be desired, cf. Ruderman 2024b . Such control with the reduced sampling or event-triggered may be advantageous in terms of the hardware components and communication architecture, but at the same time requires a reliable online estimation of the oscillation process parameters on the sensor side. The problem of online estimation of parameters from a linear regression is classical and has many solutions depending on excitation of the regressor and the properties of measurement perturbations Ljung 1987 ; Sastry and Bodson 1989 ; Astolfi et al. 2008 . A usual standing assumption requires the persistence of excitation ( Morgan and Narendra 1977 ) of the regressor for robust estimation in face of disturbances, while in the noise-free setting the necessary and sufficient condition is the interval excitation. The most popular solutions include the least squares algorithm or the gradient descent one, which guarantee an exponential convergence of the estimates to their ideal values under a sufficiently excited regressor ( Narendra and Annaswamy 1987 ), while for a weaker excitation the asymptotic convergence rates can be recovered, see Efimov et al. 2019 . In the case when some parameters appear in a nonlinear fashion in the regression, there are methods oriented on estimation for convex/concave functions of the parameters ( Fomin et al. 1981 ) or other special scenarios ( Tyukin et al. 2007 ; Ortega et al. 2024 ). Recent approaches use some preliminary algebraic operations reducing the problem to a linear regression with respect to other parameters, see e.g. Ushirobira and Efimov 2023 . A notable example here is the problem of estimation of a frequency in a measured biased harmonic signal, which has many solutions Pin et al. 2019 ; Vediakova et al. 2020 ; Ahmed et al. 2022 ; Ruderman 2024a , to mention a few. If the estimation algorithm represents a part with online control loop, then often stricter requirements on the rate of convergence and robustness are imposed. It has been observed, that in conventional approaches it is difficult to accelerate since their performance characteristics are governed by the level of excitation, Efimov and Fradkov 2015 , which may be hard to change. Recently, several alternative solutions have been proposed, which however require more computational power Chowdhary et al. 2012 ; Aranovskiy et al. 2017 , but they can provide finite/fixed-time convergence rates ( Efimov and Polyakov 2021 ) for the estimation errors in linear regressions ( Ríos et al. 2017 ; Wang et al. 2020b ). An example of the practical control scenario necessitating a rapid and reliable frequency estimation is the problem of compensating mechanical oscillations of a passive inertial load, that was recently addressed in Ruderman 2024a ; Ruderman 2024b . Based thereupon, the present work combines the power-based control ( Ruderman 2024b ) with the finite-time estimation of the biased harmonics ( Ahmed et al. 2022 ). The control scheme requires the knowledge of the instantaneous frequency, amplitude, and bias parameters of the output harmonic signal to be compensated. The rest of the paper is organized as follows. In section 2 , we resume the discrete power-based control ( Ruderman 2024b ) and provides the impulse weighting factor by analytic calculation of the amplitude reduction through the feed-forward sub-dynamics of the plant. The estimation of biased harmonics ( Ahmed et al. 2022 ) is summarized in section 3 , while its convergence prosperities and parameters tuning are addressed in section 4 . A detailed experimental case study of the oscillations compensation with online harmonics estimation is shown in section 5 . The conclusions are derived at the end in section 6 . 2 Power-based oscillations control The power-based oscillations control ( Ruderman 2024b ) is applicable as a plug-in compensator of the oscillatory output quantity y(t) , see Fig. 1 , while the overall plant transfer function G(s) assumes the double integrator in series at the output channel. The corresponding dominant conjugate-complex pole pair at \lambda_{1,2}=\sigma\pm j\omega is assumed to be either undamped i.e. \sigma=0 or very low damped i.e. 0\leq\zeta\ll 1 ; \zeta=-\sigma(\sigma^{2}+\omega^{2})^{-1/2} is the standard (normalized) damping ratio. Note that the power-based control is also capable to compensate for even slowly diverging output oscillations, i.e. for -1\ll\zeta<0 . This case will be demonstrated in the experimental study where a proportional-integral (PI) feedback controller \tilde{u} , cf. Fig. 1 , forces the otherwise stable G(s) to have the closed-loop system with a dominant pole pair with \sigma>0 . Figure 1: Discrete power-based control u^{\prime} subject to the gain shaping and time-delay synchronization. It is assumed that the single measured output of the process has the form y(t)=Y_{0}+A\sin(\omega t+\phi)+v(t),\quad t\geq 0, (1) where y(t),v(t)\in\mathbb{R} are the measured signal with the respective bounded noise. A>0 and \omega>0 are the amplitude and the frequency of oscillations, correspondingly, Y_{0}\in\mathbb{R} is the bias and \phi\in[0,2\pi) is the phase. For the purpose of estimation, we assume that all these parameters are quasi-constant omitting their explicit dependence on time in the notation. Also we note that for the power-based compensation of the harmonic part in y(t) , the estimation of the instantaneous triple [\hat{Y}_{0},\hat{A},\hat{\omega}](t) is assumed, while the phase parameter

Read the complete text