Aliased Time-Modulated Array OFDM System

Marcin Wachowiak, André Bourdoux, Sofie Pollin · 2025

Read Aliased Time-Modulated Array OFDM System by Marcin Wachowiak, André Bourdoux, Sofie Pollin (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-2503-07180.txt.

Introduction

Marcin Wachowiak and co-authors redesign a switched antenna array for OFDM signals in a 2025 paper. Time-modulated arrays steer beams by switching antennas in timed patterns, a cheap trick with two flaws. Switching must run very fast, and the pattern sprays unwanted harmonic copies of the signal.

Their fix shapes the transmitted blocks so the harmonic copies fold back onto the signal itself and add up usefully. Switching can then run slower by a chosen factor, at the cost of the same factor in data capacity. Sideband radiation drops as well. The trade suits sensing and low-rate links where hardware simplicity beats raw speed.

Catalog record

CREATORMarcin Wachowiak, André Bourdoux, Sofie Pollin
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE30 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2503-07180

Browse the full library · About the Tower of Babel

Full text

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

Aliased Time-Modulated Array OFDM System Marcin Wachowiak, André Bourdoux, Sofie Pollin Abstract The time-modulated array is a simple array architecture in which each antenna is connected to an RF switch that serves as a modulator. The phase shift is achieved by digitally controlling the relative delay between the periodic modulating sequences of the antennas. The practical use of this architecture is limited by two factors. First, the switching frequency is high, as it must be a multiple of the sampling frequency. Second, the discrete modulating sequence introduces undesired harmonic replicas of the signal with non-negligible power. In this paper, aliasing is exploited to simultaneously reduce sideband radiation and switching frequency. To facilitate coherent combining of aliased signal blocks, the transmit signal has a repeated block structure in the frequency domain. As a result, a factor A A reduction in switching frequency is achieved at the cost of a factor A A reduction in communication capacity. Doubling A A reduces sideband radiation by around 2.9 dB. I Introduction I-A Problem Statement Antenna arrays are an essential component of current and future generations of wireless systems [ 1 ] . However, when considering conventional architectures, the scaling of antenna arrays raises concerns about cost, complexity, and energy efficiency [ 2 ] , [ 3 ] . To address scaling problems, alternative array architectures should be considered [ 4 ] . A time-modulated array is a simple antenna array architecture in which each antenna is connected to the multi-throw switch that performs time modulation [ 5 ] . The idea was first proposed in [ 6 ] . To improve the efficiency of time modulation, the switch can be replaced with a low-resolution phase shifter that quickly switches [ 7 ] . The simplicity of the switch, which is the main building block of the front end, offers savings in cost, power, and size of the radio front end compared to mixers or high resolution phase shifters [ 8 ] , [ 9 ] . However, the simple architecture of TMA comes with a few limitations that prevent their widespread adoption. Firstly, in conventional time-modulated arrays, the switch must operate at a multiple of the sampling frequency to take advantage of oversampling to offer an improved phase-shifting resolution and prevent the overlap of the harmonics of the transmitted signal [ 5 ] , [ 10 ] . Secondly, due to the discretized modulating sequence, the time modulation introduces harmonic replicas of non-negligible power of the baseband signal at multiples of the switching frequencies. [ 7 ] , [ 11 ] . To comply with wireless standards, the sideband radiation must be suppressed, which requires an additional bandpass filter per antenna. To facilitate more widespread utilization of TMAs, methods are needed to reduce the switching frequency and suppress the sideband radiation. The harmonic components might be exploited to facilitate the simultaneous multi-beam operation with TMA [ 12 ] [ 13 ] . However, each harmonic beam transmits the same replicated baseband signal, limiting the transmit TMA’s practical application. Fig. 1: Architecture of the aliased time-modulated transmit array. The baseband signal consists of A identical repeated blocks in the frequency domain. Each block is differently precoded by v(a,d) , depending on the block index a and selected delay value d . The oversampling factor O_{\tau} determines the number of required frequency precoders implemented as digital baseband units. Depending on the index d , each m -th antenna is fed by a selected precoder via a distribution network of standard RF switches. I-B Relevant works Initially, time-modulated arrays were analyzed under a narrowband scenario considering just a single carrier. Various algorithms have been proposed to optimise the switching instants of the RF switches to minimise sideband radiation, including genetic [ 14 ] , evolution [ 15 ] , [ 16 ] and particle swarm [ 17 ] optimization algorithms. The modulating signal format determines the sideband radiation. To further reduce sideband radiation, the following approaches have been proposed: aperiodic modulating pulses [ 18 ] , pseudorandom modulating sequence [ 19 ] or different switching frequency per element [ 20 ] . A great improvement in the sideband suppression is observed when employing phase modulation [ 21 ] , utilizing stepped waveforms [ 22 ] or simultaneous amplitude and phase modulation (IQ) [ 23 ] , [ 24 ] , [ 25 ] . The interest in the TMA architecture has resulted in practical implementations which utilize simple low-bit switched delay lines as building blocks [ 26 ] , [ 27 ] . The listed approaches are suitable for narrowband signals, and the switching frequency is assumed to be much higher than the transmitted signal bandwidth. This leads to an assumption of an arbitrary and unconstrained timing resolution of a switch. For wideband signals, this assumption is no longer valid and the switching frequency becomes constrained by signal bandwidth and quickly approaches limits imposed by hardware. Moreover, the extent of the sideband radiation is increased as the baseband signal now has a greater width in the frequency domain, increasing the strain on filtering requirements. I-C Contributions his work considers the TMA combined with a wideband OFDM signal and proposes a method to simultaneously reduce the switching frequency and the sideband radiation by introducing aliasing between the harmonic replicas of the transmitted signal. It is achieved at the cost of the effective bandwidth as the baseband signal, in the frequency domain, is composed of A repeated blocks, which are differently precoded. The improvement in sideband suppression and spectral efficiency trade-off are discussed in detail. The oversampling of the modulating sequence for improved resolution beamforming with aliased TMA is proposed and studied. II Signal model II-A Time modulation Consider a time-modulated antenna, which is connected to the N throw/state switch. The RF switch acts as a discrete phase shifter - each state of the switch results in a different phase shift of the input signal. The phase shift of n -th switch state is given by \varphi(n)=2\pi\frac{n}{N}, (1) where n is the index of the state n\in\{0,1,\ldots,N-1\} . The switch operates at a frequency f_{\mathrm{s}} . The pulse duration of each state is T_{\mathrm{p}}=AT_{\mathrm{s}},\ A\in\mathbb{N}^{+}, (2) where A is the aliasing factor that increases the pulse duration and reduces pulse frequency. The pulse frequency is f_{\mathrm{p}}=\frac{1}{T_{\mathrm{p}}}=\frac{f_{\mathrm{s}}}{A}. (3) The modulating signal is a periodically repeated sequence of N rectangular pulses with a phase given by ( 1 ), for the case of linear phase steps, which results in maximization of the power at the first harmonic [ 7 ] . Due to the discretized switching instants and phase shift values, the basic building block of the TMA modulating signal is a rectangular pulse. Consider a periodical rectangular pulse of duration T_{\mathrm{p}} with unit amplitude and phase shift of \varphi(n) . The time domain representation of the pulse is given by g_{n}(t)=e^{j\varphi(n)}\sum_{i=-\infty}^{\infty}\rect{\left(\frac{t}{T_{\mathrm{p}}}-\frac{T_{\mathrm{p}}}{2}+iNT_{\mathrm{p}}\right)}. (4) The pulse is periodical with a period NT_{\mathrm{p}} which allows to expand it into a Fourier series. The Fourier coefficient of the k -th harmonic component is given by \displaystyle G_{n}(k) \displaystyle=e^{j\varphi(n)}\sinc{\left(\pi\frac{k}{N}\right)}e^{-j\pi\frac{k}{N}}. (5) The modulating (control) signal c(t) is composed of a sequence of N delayed complex pulses from ( 4 ) resulting in \displaystyle c(t) \displaystyle=\sum_{n=0}^{N-1}g_{n}\left(t-nT_{\mathrm{p}}\right). (6)

Read the complete text