Hostname: page-component-cd9895bd7-fscjk Total loading time: 0 Render date: 2024-12-28T17:51:38.624Z Has data issue: false hasContentIssue false

Multi-pattern synthesis in fourth-dimensional antenna arrays using BGM-based quasi-Newton memetic optimization method

Published online by Cambridge University Press:  01 August 2023

Avishek Chakraborty*
Affiliation:
Department of EECE, GST, GITAM University, Bengaluru, India
Ravi Shankar Saxena
Affiliation:
Department of ECE, GMR Institute of Technology Rajam, Andhra Pradesh, India
Anshoo Verma
Affiliation:
Department of CE, IES Institute of Technology and Management, IES University, Bhopal, India
Ashima Juyal
Affiliation:
Uttaranchal Institute of Technology, Uttaranchal University, Dehradun, India
Sumit Gupta
Affiliation:
Department of ECE, SR University, Warangal, Telangana, India
Indrasen Singh
Affiliation:
School of Electronics Engineering, VIT Vellore, Tamil Nadu, India
Gopi Ram
Affiliation:
Department of ECE, NIT Warangal, Telangana, India
Durbadal Mandal
Affiliation:
Department of ECE, NIT Durgapur, West Bengal, India
*
Corresponding author: Avishek Chakraborty; Email: avishekdreamz@gmail.com
Rights & Permissions [Opens in a new window]

Abstract

The advancement in wireless communication is fueling the growth of innovative antenna array designs toward cost-effective and performance-oriented solutions. This paper proposed unconventional methods to design antenna arrays for multi-pattern synthesis without using attenuators or phase shifters. A low-cost alternative is proposed with “Time-modulation”-based antenna array capable of electronic scanning and beam steering. Here, “Time” is utilized as a fourth-dimensional (4D) array parameter, and that is why “Time-modulated” arrays are also called as 4D antenna arrays. The idea is to control the high-speed switch attached with each antenna periodically to produce desirable current and phase tapering. This article expanded the “Time-modulation” concept to synthesize multiple radiation patterns like monopulse patterns, scanned beam patterns, shaped beam patterns, and cosecant-squared beam patterns for multifunctional radar systems. Suitable time schemes are developed to generate the narrowband sum–difference patterns useful for monopulse radars. Simultaneous scanned beam patterns are also proposed for narrowband communication. Furthermore, to address the wideband applications, shaped flat-top beam patterns and cosecant-squared beam patterns are also proposed. In this regard, 20- and 16-element “Time-modulated” linear array antennas are developed, and the parameters of the arrays are controlled by suitably designed objective functions with quasi-Newton method (QNM)-based memetic optimization method. For this purpose, first a well-known genetic algorithm is adopted to search the potential trust regions in the exploration stage and QNM is used for fine-tuning. Furthermore, the Broyden’s good method-based direction-updating equation is used with QNM to improve the performance.

Type
Antenna Design, Modelling and Measurements
Copyright
© The Author(s), 2023. Published by Cambridge University Press in association with the European Microwave Association

Introduction

Array antennas are the fundamental technologies of a modern communication system that has a broad range of applications in radar, remote sensing, satellite navigation, biomedical imaging, microwaves, and several other fields [Reference Hansen1]. Since its introduction, antenna arrays have evolved substantially [Reference Mailloux2] and incorporated to multi-domain research prospect related to electromagnetics, electronics, computer science, control systems, etc. [Reference Chakraborty, Ram and Mandal3]. Phased arrays are the advancement over traditional antenna arrays usually considered for electronic beam scanning and beam steering [Reference Rocca, Oliveri, Mailloux and Massa4]. Generally, the phased array antennas are equipped with transmit and receive modules that permit the independent control of amplitude and phase of the transmitted or received signals with the help of attenuators and phase shifters [Reference Rocca, Oliveri, Mailloux and Massa4]. Cutting-edge communication system has thrusted stringent stipulation on phased arrays like multi-functionalities, reconfigurabilities, lesser weight, more compact, and reduced costs [Reference Rocca, Oliveri, Mailloux and Massa4]. Further, the requirement of beamforming, beam scanning, and beam steering in advance communication systems has significantly increased the research concern in the domain of phased arrays. So, the exploration and exploitation of newer techniques for alternative solution is needed.

An unorthodox alternative to phased antennas has got the attention of antenna research community when Shanks et al. introduced the concept of fourth-dimension—“Time” to control the radiation patterns of antenna arrays [Reference Shanks and Bickmore5]. Immediately after their path-breaking research, Shanks proposed a method for electronic scanning by utilizing the idea of “Time-modulation” [Reference Shanks6]. Although the “Time-modulated” array (TMA) was initially postulated for achieving ultra-low sidelobe radiation patterns [Reference Kummer, Villeneuve, Fong and Terrio7], eventually they have become a popular choice for several other applications like electronic beam scanning [Reference Haupt8], beam shaping [Reference Yang, Chen and Nie9], beam steering [Reference Chakraborty, Ram and Mandal10], and so on. TMAs may be classified as the “Time-domain” counterparts of traditional antenna arrays where a synchronized sequence is implemented with swiftly turning ON and OFF the RF switches, attached with each antenna element [Reference Rocca, Yang, Poli and Yang11]. By adding and removing the elements from beamformer network (BFN), a desirable time-averaged radiation pattern can be generated [Reference Maneiro-Catoira, Brégains, García-Naya and Castedo12]. The turn-ON period and switch-ON instants of each radiators can be controlled to obtain a sequence equivalent to the current and phase tapering in conventional arrays [Reference Maneiro-Catoira, Brégains, García-Naya and Castedo12]. Hence, the need of costly phase shifters and erroneous attenuators may be eradicated, and the quantization error and insertion loss that occurred in traditional arrays may be avoided [Reference Fondevila, Brégains, Ares and Moreno13]. However, the added advantages of TMAs are also accompanied by the inherent drawback of infinite number of harmonic pattern generation due to the periodicity of the switches [Reference Yang, Gan and Tan14]. Harmonic radiations are generally considered unwanted because a major part of the dissipated power is transferred from the main beam to the sidebands [Reference Brégains, Fondevila-Gómez, Franceschetti and Ares15]. The sideband radiation (SR) should be reduced to improve the total efficiency of the TMAs [Reference Yang, Gan and Qing16]. The primary research work in this field was focused only on sidelobe levels (SLLs) and SR reduction [Reference Yang, Can and Tan17]. However, in due course, a new dimension has evolved with the notion of exploiting the unwanted SRs as an advantage. In this regard, TMAs for direction finding has been introduced [Reference Tennant and Chambers18]. The possibilities have been enhanced after Li et al. proposed a technique to steer the beam in some specific directions [Reference Li, Yang, Chen and Nie19]. From then on, various research works have been reported with efficiently steered and scanned performances [Reference Tong and Tennant20Reference Ram, Panduro, Reyna, Kar and Mandal25]. This paper exploited the notion further by proposing efficient multi-patterns such as sum–difference pattern for monopulse radars and scanned beam and shaped beam patterns for multifunctional radars (MFRs).

Monopulse-based searching and tracking radar require an array system that can produce a sum and a difference pattern simultaneously [Reference Barton26]. Tracking radar automatically keeps the main array beam aligned to the target object, producing a narrow and highly directive radiation pattern [Reference Sherman27]. Monopulse arrays have been broadly employed in radar tracking systems to improve the poor speed and complicated BFNs experienced by a conical or sequential scanning [Reference Sherman27]. The spatially independent and simultaneously generated harmonic beams can be used with appropriate switching sequence to deal with the drawbacks of compromised sum–difference patterns [Reference Chakraborty, Ram and Mandal28]. This article comes up with a suitable optimized switch scheme to produce the simultaneous monopulse pattern (sum–difference pattern) in a 20-element time-modulated linear array (TMLA). Further research on electronic beam scanning beneficial for the MFRs has been addressed by designing the suitable optimal time schemes [Reference Chakraborty, Ram and Mandal29]. Steering the consecutive harmonic patterns to prespecified angles and simultaneously lowering the SLLs is extremely beneficial for radars [Reference Chakraborty, Ram and Mandal30Reference Chakraborty, Ram and Mandal32]. In this article, a 20-element TMLA is used to produce the scanned harmonic pattern at ±30° from the broadside. Furthermore, a shaped flat-top pattern is also targeted with a 16-element TMLA. To obtain the desired optimal radiation patterns, a memetic quasi-Newton optimization algorithm is employed. Here the limitations of traditional Newton method are eliminated by incorporating a Broyden’s good method (BGM)-based direction-updating equation in the exploitation stage, along with the advantages of finding potential trust regions at the exploration stage using genetic algorithm (GA).

The work proposed in this article have manifold prospects briefly pointed out as follows:

  • First of all, multiple patterns have been addressed by exploiting the concept of time-modulation.

  • Second, for all the beam patterns, interference rejection is considered of the utmost importance.

  • Third, to cancel the interference from unwanted sources, SLLs of all the patterns are minimized.

  • Fourth, the unwanted higher-order harmonics apart from the first sideband are all suppressed.

  • Fifth, the arrays are designed to cater the need of both the narrowband beam patterns using scanned beams and monopulse beams, as well as the wideband patterns using flat-top and cosecant beams.

The rest of the article is described as follows: The theoretical and mathematical insights of TMLA are presented in Section 2. The memetic algorithm used to optimize the control parameters is briefly discussed in Section 3. In Section 4, the outcomes of the proposed methods are analyzed, and a comparison and thorough discussion on future prospect is furnished. Finally, conclusive remarks are given in Section 5.

Theoretical background

Time-modulated linear arrays

A TMLA consists of N isotropic radiating antennas attached with same number of Single-Pole-Single-Through (SPST) Radio-Frequency (RF) switches, and it is controlled by a complex programmable logic device (CPLD), which is exhibited in Fig. 1. The array factor (AF) for the same may be given as [Reference Chakraborty, Ram and Mandal31]

(1)\begin{equation}{\textrm{AF}}\left( {\theta ,t} \right) = {{\textrm{e}}^{j\left( {2\pi {f_0}} \right)t}} \sum\limits_{n = 1}^N {I_n}{U_n}\left( t \right){{\textrm{e}}^{jk\left( {n - 1} \right)d\,\cos \theta }}\end{equation}

Figure 1. N-element array connected through switches and managed by a CPLD.

where In denotes the amplitude of the nth antenna, k is the constant of propagation, d represents the uniform spacing between antennas, θ is the angle of the impinging signal, f 0 is the operating frequency with T 0 as the period, and Un(t) is the time-switching function.

Due to periodicity of switches, Un(t) can be given as [Reference Chakraborty, Ram and Mandal32]

(2)\begin{equation}{U_n}\left( t \right) = \sum \limits_{m = - \infty }^\infty {a_{mn}}{{\textrm{e}}^{jm\left( {2\pi {f_p}} \right)t}}\end{equation}

The Fourier coefficient (amn) of nth element for mth frequency term, inherently generated because of time-modulation, can be presented as [Reference Bhattacharya, Saha and Bhattacharyya33]

(3)\begin{equation}{a_{mn}} = {1 \over {{T_p}}} \smallint \limits_0^{{T_p}} {U_n}\left( t \right){{\textrm{e}}^{ - jm\left( {2\pi {f_p}} \right)t}}\,{\textrm{d}}t \end{equation}

Hence, the AF of uniformly excited (In = 1) TMLA can be expressed as

(4)\begin{equation}{\textrm{AF}}\left( {\theta ,t} \right) = \sum \limits_{m = - \infty }^\infty \sum \limits_{n = 1}^N {a_{mn}}\left\{ {{{\textrm{e}}^{jk\left( {n - 1} \right)d\,\cos \theta }}} \right\}{{\textrm{e}}^{j2\pi \left( {{f_0} + m{f_p}} \right)t}}\end{equation}

The array pattern at mth harmonics may be simplified as [Reference Chakraborty, Ram and Mandal34]

(5)\begin{equation}{\textrm{A}}{{\textrm{F}}_m}\left( {\theta ,t} \right) = {{\textrm{e}}^{j2\pi \left( {{f_0} + m{f_p}} \right)t}} \sum \limits_{n = 1}^N {a_{mn}}{{\textrm{e}}^{jk\left( {n - 1} \right)d\,\cos \theta }}\end{equation}

The main beam pattern can be represented with m = 0, and SR patterns are at the multitude of modulation frequencies (mfp) where m = ±1, ±2, ±3, …, ±∞.

Time sequences for diverse applications of TMLAs may be designed, and to explore that, harmonic beam steering in TMLA is inscribed in the paper by suitably optimized shifted time sequences. The general ON–OFF time sequence for the nth antenna is given in Fig. 2(a), in which the radiator is turned on for the duration τn (0 ≤ τnTp). The pulse initiated at $\tau _n^1 = 0$ and terminated at $\tau _n^2$ having a normalized ON-time duration $\{ \left( {\tau _n^2 - \tau _n^1} \right)/{T_p}\} $. The switching function and the Fourier coefficient are given as [Reference Chakraborty, Ram and Mandal34]

(6)\begin{equation}{U_n}\left( t \right) = \left\{ \begin{matrix} {1, \tau _n^1 \le t \le \tau _n^2 \le {T_p}\ {\textrm{where}}\ \tau _n^1 = 0 } \\ {0, {\textrm{otherwise}}} \\ \end{matrix}\right.{\textrm{ }}\end{equation}
(7)\begin{equation}{a_{mn}} = {{{\tau _n}} \over {{T_p}}}\left\{ {{\textrm{sinc}}\left( {m\pi {f_p}{\tau _n}} \right)} \right\}{{\textrm{e}}^{ - jm\pi {f_p}\left( {{\tau _n}} \right)}}\end{equation}

Figure 2. (a) Simple switching sequence and (b) shifted switching sequence for nth radiator in TMLA.

The generalized shifted scheme of the nth antenna is given in Fig. 2(b), in which the ON-time duration of Fig. 2(a) is moved to $\tau _n^1$ ($\tau _n^1 \ne 0$) and the nth antenna remains ON up to $\tau _n^2$. The switching function and excitation coefficient for the shifted sequence can be given as

(8)\begin{equation}{U_n}\left( t \right) = \left\{ \begin{matrix} {1, 0 \lt \tau _n^1 \le t \le \tau _n^2 \le {T_p}\ {\textrm{where}}\ \tau _n^1 \ne 0 } \\ {0, {\textrm{otherwise}} } \\ \end{matrix} \right.{\textrm{ }}\end{equation}
(9)\begin{equation}{a_{mn}} = {{\left( {\tau _n^2 - \tau _n^1} \right)} \over {{T_p}}}{\ }\left[ {{\textrm{sinc}}\left\{ {m\pi {f_p}\left( {\tau _n^2 - \tau _n^1} \right)} \right\}} \right]{\ }{{\textrm{e}}^{ - jm\pi {f_p}\left( {\tau _n^1 + \tau _n^2} \right)}}\end{equation}

For steering the first sideband patterns (m = ±1) in prespecified direction (${\theta _0}$), the switching instants for each array element can be modified as [Reference Rocca, Zhu, Bekele, Yang and Massa35]

(10)\begin{equation}\tau _n^1 = \left[ {{{\left( {n - 1} \right)kd\ {\textrm{cos}{\theta _0}}} \over {2\pi }} - {{{\tau _n}} \over 2}} \right]{\textrm{ mod }}1{\textrm{ }}\end{equation}

The amount of radiated power for the fundamental pattern (P 0), the total power dissipated by harmonic patterns including the fundamental one (P T), and the directivity can be given as [Reference Poddar, Paul, Chakraborty, Ram and Mandal36]

(11)\begin{equation}{P_0} = \smallint\limits_{2\pi }^0 \smallint\limits_\pi ^0 {\left| {A{F_0}\left( {\theta ,\phi } \right)} \right|^2}\sin \theta {\textrm{d}}\theta {\textrm{d}}\phi \end{equation}
(12)\begin{equation}{P_T} = \sum\limits_\infty ^{m = - \infty } \smallint\limits_{2\pi }^0 \smallint\limits_\pi ^0 {\left| {A{F_m}\left( {\theta ,\phi } \right)} \right|^2}\sin \theta {\textrm{d}}\theta {\textrm{d}}\phi \end{equation}
(13)\begin{equation}D = {{4\pi {{\left| {A{F_0}\left( {{\theta _0},{\phi _0}} \right)} \right|}^2}} \over {\sum\nolimits_{m = - \infty }^\infty \smallint\nolimits_0^{2\pi }\smallint\nolimits_0^\pi {{\left| {A{F_m}\left( {\theta ,\phi } \right)} \right|}^2}\sin \theta {\textrm{d}}\theta {\textrm{d}}\phi }}\end{equation}
where the central beam pattern (at m = 0) points toward θ = θ 0, ϕ = ϕ 0 and AFm (θ, ϕ) indicates the beam patterns generated at the mth-order harmonic frequencies (m ≠ 0).

Objective functions

To manage the radiation characteristics of the central beam pattern and the steered SR patterns, the SLLs are lowered. The undesired higher sidebands are minimized. To obtain the coveted outcomes, an appropriate objective function (OF1) is designed as

(14)\begin{equation}{\textrm{O}}{{\textrm{F}}_{\textrm{1}}} = {\left. {({\textrm{SLL}}_0^{\left( i \right)})} \right|_{{f_0}}} + {\left. {({\textrm{SLL}}_1^{\left( i \right)})} \right|_{{f_0} \pm {f_p}}} + {\ }({\left. {{\textrm{SBL}}_m^{\left( i \right)})} \right|_{{f_0} + m{f_p}}}{\textrm{ }}\end{equation}

where i denotes the current iteration, SLL0 and SLL1 represents the maximum SLLs at the fundamental (f 0) and first harmonic frequencies (${f_0} \pm {f_p}$), and SBLm is the maximum level of sideband patterns (${f_0} + m{f_p}$ for m = ±2, ±3, …, ±∞).

The exploitation of the harmonic radiations is concerned of highly directive radiation pattern generation for narrowband communication. On the other hand, the shaped pattern synthesis is aimed toward designing a flat-top pattern and the cosecant-squared pattern at the first sideband for specific wideband applications. The problems are cast into suitable OFs (OF2 for flat-top and OF3 for cosecant-squared patterns) as

(15)\begin{align}{\textrm{O}}{{\textrm{F}}_{\textrm{2}}} & = ({\textrm{SLL}}_{{\textrm{max}}}^{\left( i \right)}){|_{{f_0}}} + ({\textrm{Ripple}}_{{\textrm{max}}}^{\left( i \right)}){|_{{f_0} + {f_p}}}\nonumber \\ & \qquad + {{\left( {\beta _{{\textrm{Trans}}}^{\left( i \right)}} \right)} \over {{90}^{\circ}}}{|_{{f_0} + {f_p}}}{\textrm{ + }}({\textrm{SBL}}_m^{\left( i \right)}){|_{{f_0} + m{f_p}}}{\textrm{ }}\end{align}
(16)\begin{align}{\textrm{O}}{{\textrm{F}}_{\textrm{3}}} & = ({\textrm{SLL}}_{{\textrm{max}}}^{\left( i \right)}){|_{{f_0}}} + ({\textrm{Ripple}}_{{\textrm{max}}}^{\left( i \right)}){|_{{f_0} + {f_p}}}\nonumber \\ & \qquad + {{\left( {\beta _{{\textrm{Trans}}}^{\left( i \right)}} \right)} \over {{90}^{\circ}}}{|_{{f_0} + {f_p}}}{\textrm{ + }}(\beta _C^{\left( i \right)} - {\beta _0}){|_{{f_0} + {f_p}}}{\textrm{ }}\end{align}
where Ripplemax is the maximum ripple calculated at the first sideband, β Trans is the transition width of the first sideband pattern, βC is the calculated direction of the maximum power radiation, β 0 is the direction of maximum radiation in the desired plane. To minimize the OF1, OF2, and OF3, a quasi-Newton method (QNM)-based memetic optimization technique is employed.

Quasi-Newton memetic optimization

Evolutionary optimization methods are population-based computational algorithms that help to unfold real-world technical issues by imitating the behavioral orientations of natural phenomenon [Reference De Jong, Fogel, Schwefel, Baeck, Fogel and Michalewicz37]. Bioinspired computation methods have evolved as suitable candidates for solving electromagnetics problems [Reference Del Ser, Osaba, Molina, Yang, Salcedo-Sanz, Camacho, Das, Suganthan, Coello and Herrera38]. Researchers have found these algorithms useful for synthesizing the radiation patterns of linear arrays [Reference Ram, Mandal, Kar and Ghoshal39, Reference Durmus and Kurban40], circular arrays [Reference Rattan, Patterh and Sohi41, Reference Das, Mandal and Kar42], planar arrays [Reference Dib43, Reference Bogdan, Godziszewski and Yashchyshyn44], concentric circular arrays [Reference Dib and Sharaqa45, Reference Ram, Mandal, Kar and Ghoshal46], hexagonal arrays [Reference Misra and Mahanti47], etc. Memetic algorithms are the population-based metaheuristics techniques, where the strengths of global search and the fine-tuning properties of local search techniques are combined [Reference Neri and Cotta48]. GA is a familiar global search, stochastic optimization method based on natural selection and evolution [Reference Weile and Michielssen49]. GA has an excellent global search ability but requires a higher number of iterations for convergence [Reference Weile and Michielssen49]. On the other hand, deterministic QNM is the advanced version of the Newton method where a modified Hessian matrix is incorporated to minimize the computation cost and complexities of the Newton method. The hybridization of both can be regarded as a memetic computation method for solving nonlinear optimization problems [Reference Zhang, Lin, Zhang and Li50]. In this paper, QNM-based memetic GA (QNM-GA) is employed.

The genetic algorithm

GA is a kind of probabilistic search method that utilizes the idea of natural selection and evolution. At each iteration, it preserves a population of individuals, also called as Chromosomes, coded from possible solutions (xk) such as [Reference Zhang, Lin, Zhang and Li50]:

(17)\begin{equation}{x_k} = {x_{{\textrm{min}}}} + {r_k}\left( {{x_{{\textrm{max}}}} - {x_{{\textrm{min}}}}} \right)\end{equation}

where x min and x max are the lower and upper boundaries of the search space and rk denotes some positive random values within the range of 0 and 1. Chromosomes are constructed over the binary range [0, 1] so that the Chromosome values can be uniquely mapped onto the decision variable domain. Each Chromosome is then evaluated by the fitness function, which is nothing but the representation of the optimization problem in hand. Based on that, elite strings are generated and crossover and mutation operations are done hereafter to generate the offsprings. In crossover operation, GA makes use of the following equation to produce two parent solutions xs and xt based on the crossover rate Cr as [Reference Zhang, Lin, Zhang and Li50]:

(18)\begin{equation}{x_{kj}} = \left\{ \begin{matrix} {{x_{sj}}, {\textrm{if}}\ {C_r} \lt {r_{j0}} } \\ {{x_{tj}}, {\textrm{otherwise}}} \\ \end{matrix} \right.{\textrm{ }}\end{equation}

where rj 0 is a random number within the range of 0 and 1. Then the mutation operation is done based on mutation rate parameter Pm as [Reference Zhang, Lin, Zhang and Li50]:

(19)\begin{equation}{x_{kj}} = \left\{ \begin{matrix} {{x_{kj}} + {r_{j2}}\left( {{x_{kj}} - {x_{mj}}} \right), {\textrm{if}}\ {P_m} \lt {r_{j1}} } \\ {{x_{kj}}, {\textrm{otherwise}}} \\ \end{matrix} \right.\end{equation}

where rj 1 and rj 2 are the random numbers within the range of 0 and 1. In this way, the genetic cycle updating continues iteratively until the completion of exploration stage for finding out the potential trust regions.

BGM-based QNM

Newton’s method is an iterative method often used to solve the nonlinear equations due to its capability of rapid convergence from a sufficiently good starting position [Reference Li, Qi and Roshchina51]. The limitation of Newton’s method is the direct computation of Jacobian, which is computationally expensive [Reference Li, Qi and Roshchina51]. In addition, the Newton’s method is not suitable when the size is too large [Reference Zhou and Zhang52]. On the other hand, QNM is an updated version of Newton’s method where the computation of Jacobian is skipped without affecting the locally super linear convergence characteristics [Reference Zhou and Zhang52]. QNM is based on the gradient method with double differentiation, implying better local search capability and faster convergence characteristic [Reference Fang, Ni and Zeng53]. However, double differentials of antenna array are not possible as it is a simulation-based large-scale model [Reference Fang, Ni and Zeng53]. Hence, a derivative-free QNM is adopted where BGM is used for updating the trust regions to get the best solutions for the problems in hand [Reference Fang, Ni and Zeng53]. The brief overview of BGM-based QNM is discussed below:

Let us consider a set of nonlinear equations: $F\left( x \right) = 0, x \epsilon {\ }{\Re ^n}$, where $F:{\ }{\Re ^n} \to {\ }{\Re ^n}$ is continuously differentiable and its Jacobian is denoted by (${J_{{x_k}}}$). Here, ${x_k}$ is the vector of variables such that $F\left( {{x_k}} \right)$ is the best cost value of the objective function at kth iteration. In a conventional Newton’s method, calculation of Hessian matrix and the inverse of it is computationally expensive, especially when the dimensions are large. The Hessian of function $F:{\ }{\Re ^n} \to {\ }{\Re ^n}$ is the Jacobian of its gradient. In QNM, the Hessian is approximated by positive definite n × n symmetric matrix ${B_k}$ that is updated iteratively such that the Hessian approximate (${B_k}$) of Jacobian (${J_{{x_k}}}$) must satisfy the quasi-Newton condition (secant equation) [Reference Li, Qi and Roshchina51]:

(20)\begin{equation}{B_{k + 1}}\left( {{x_{k + 1}} - {x_k}} \right) = \nabla F\left( {{x_{k + 1}}{\ }} \right) - \nabla F\left( {{x_k}{\ }} \right){\textrm{ }}\end{equation}

where the gradient vector is obtained from the first-order Taylor expansion of $\nabla F\left( {{x_{k + 1}}{\ }} \right)$ about $\nabla F\left( {{x_k}{\ }} \right)$ computed by finite difference method (FDM). The quasi-Newton condition of eq. 20 can be rewritten more succinctly by letting $\nabla F\left( {{x_{k + 1}}{\ }} \right) - \nabla F\left( {{x_k}{\ }} \right) = {y_k}$ and $\left( {{x_{k + 1}} - {x_k}} \right) = {s_k}$, so that we have [Reference Zhou and Zhang52]

(21)\begin{equation}{B_{ + 1}}{s_k} = {y_k}{\textrm{ }}\end{equation}

For the derivative-free quasi-Newton equation, Broyden has further extended the updating equation, which is known as BGM. The update equation of BGM is as follows [Reference Fang, Ni and Zeng53]:

(22)\begin{equation}{B_{k + 1}} = {B_k} + {{\left( {{y_k} - {B_k}{s_k}} \right)s_k^T} \over {s_k^T{s_k}}}{\textrm{ }}\end{equation}

Finally, the derivative-free quasi-Newton equation for unconstrained optimization can be represented by incorporating the Broyden’s method and direct line search method as [Reference Krishna Chaitanya, Raju, Raju and Mallikarjuna Rao54]:

(23)\begin{equation}x_k^{{\textrm{new}}} = {x_k} + {\alpha _k}{d_k}\end{equation}

where ${\alpha _k}$ is the step size chosen to satisfy the Wolfe conditions, which is a set of inequalities used for line search in QNM. Here, ${d_k} = - {{\bf{H}}_k}{y_k}$ denotes the Broyden-based direction, ${y_k}$ is the gradient vector, and ${{\textbf{H}}_k}$ is the Hessian that is approximated by BGM.

The gradient vector ${y_k}$ in QNM is calculated using FDM, which is based on the idea of replacing derivatives with finite differences. It is not suitable for exploration as it may not be able to approximate the function derivatives [Reference Krishna Chaitanya, Raju, Raju and Mallikarjuna Rao54]. Instead of that, GA can gradually converge to global optima and can be better fitted for the exploration stage to find out the potential trust regions in a multidimensional search space. That means step size will be smaller with the iterations going on. Then FDM can be used to exploit the trust regions in later evolution stage to produce the accurate results. The memetic method can be performed with a step-by-step process starting with the initialization of GA by setting the parameter values and population size for the concerned problem. After evaluating the fitness value of each population agent, crossover and mutation operation is executed iteratively to find the best possible trust regions. Then, QNM is applied to perform the local search within the potential trust regions for a finely tuned optimal solution. The whole process is repeated until the stopping criteria are reached [Reference Krishna Chaitanya, Raju, Raju and Mallikarjuna Rao54]. The flow chart of the QNM-GA is shown in Fig. 3.

Figure 3. Flow chart of QNM-based memetic GA optimization.

Figure 4. Switching sequence obtained with QNM-GA to get the sum–difference beam pattern in a 20-element TMLA for MTR.

Numerical results

This section discusses several applications of TMLAs through distinct examples where QNM-GA method is adopted to obtain the optimized switching schemes for achieving the desired goals. The array is operating at 3 GHz (f 0) with 1 GHz of modulating frequency (fp). The uniformly excited (In = 1) TMLA with a fixed inter-element spacing (d = λ/2) has these properties: −13.2 dB SLL, 6.48° half-power beamwidth (HPBW), 14.4° first-null beamwidth (FNBW), and a directivity of 12.04 dB. The numerical outcomes are simulated with MATLAB. The best-reported parameters of QNM-GA are used for the optimization process and also reported in Table 1 [Reference Krishna Chaitanya, Raju, Raju and Mallikarjuna Rao54]. The proposed approach can also be considered with active elements using active element pattern method through pattern multiplication [Reference Morabito, Di Carlo, Di Donato, Isernia and Sorbello55, Reference Palmeri, Isernia and Morabito56]. But, for mathematical simplicity, only isotropic electromagnetic radiators are considered in this article. Further investigation is always possible by using different active elements with the proposed concept.

Table 1. Control parameters of QNM-based memetic GA optimization

Monopulse scanned beam patterns

The simultaneous sum–difference beam for monopulse scanning is targeted using a 20-element TMLA. To achieve this target, an in-phase switching scheme for the first 10 radiators and an out-of-phase time scheme for the others with respect to array phase center is implemented. The notion is to design a switching scheme so that the OFF periods of the first 10 elements shall be the ON periods for the other 10. Hence, an in-phase and a reversed-phase tapering may be initiated. As a consequence, sum and difference beam with deep null toward the main beampattern (broadside direction) may be obtained at the same time. The QNM-GA-based switching scheme for the simultaneous monopulse scanned pattern is shown in Fig. 4, and the relative power patterns acquired by the switching sequence are given in Fig. 5. The desirable aim is to produce a sum pattern at the central frequency and a squint difference beam pattern at the first harmonics.

Figure 5. The corresponding sum pattern at central frequency and the squinted difference pattern at first harmonic obtained with QNM-GA-based optimal in and out-of-phase time sequence presented in Fig. 4.

The suggested technique using a 20-element TMLA has achieved an ultra-low SLL0 of −40.89 dB at fundamental frequency. The offset (SBL) in difference beam is calculated as −11.29 dB. The SLL1 of the difference beam is obtained as −22.54 dB. The HPBW and FNBW of the relative power patterns attained at f 0 and (${f_0} \pm {f_p}$) are 7.53° and 21.59°, and 4.89° and 10.1°, respectively. Out of total dissipated power, 71.8152% is utilized to produce the desired beams and 28.1848% is dissipated in higher sidebands. Thus, an efficient, controlled, and simultaneous monopulse scanned pattern is achieved with the suggested method beneficial for monopulse tracking radars (MTR). The optimized ON times and starting times are reported in Table 2. The results and the comparison with other published results are presented in Tables 3 and 4, respectively.

Table 2. Optimized ON-time period and starting times for multi-pattern synthesis

Table 3. Obtained simulation results for multi-pattern synthesis using QNM-GA

Table 4. Comparison of the proposed work for multiple pattern generation

Note: θ 0, direction of the main beam; θ 1, scan angle for first positive harmonics; θ −1, scan angle for first negative harmonics; and NR, not reported.

Simultaneous scanned beam patterns

The abilities of TMLA are further explored by exploiting the lower sideband patterns for scanning multiple beams. The undesired SR at harmonic frequencies is utilized by suitable time scheme to acquire the scanned beam patterns in some specific directions. The estimated direction of arrival observed for this problem is ±30° from the broadside.

The QNM-GA-based optimal time scheme for ±30° scanned beam patterns are given in Fig. 6, and the corresponding radiation patterns are presented in Fig. 7. The desired objective is to obtain the scanned beam patterns at ±30° from broadside. The first negative (f 0 − fp) and first positive (f0 + fp) sideband beams are optimized to get the scanned harmonic patterns pointed at 60° (–30° shifted) and 120° (+30° shifted) while the fundamental beam at f0 is pointing toward 90°. The central frequency beam has acquired a lowered SLL0 of −33.28 dB. The SBL of first-order sideband beam (|m| = 1) is −1.45 dB. The SLL1 of −23.58 dB is also acquired for the scanning patterns comparing with the highest level (i.e., the SBL). 73.5913% of total power is utilized in the desired beams and 26.4087% is dissipated in harmonics. The radiated power in the fundamental, first positive, and negative sideband beams are calculated as 31.2099%, 21.1907%, and 21.1907% of the total power. Thus, an alternative yet less costly solution for electronic scanning is proposed where the disadvantages of traditional phased arrays are eradicated. The optimized ON times and starting times are presented in Table 2. The results and the comparison with other published results are presented in Tables 3 and 4, respectively.

Figure 6. Switching sequence obtained with QNM-GA to generate scanned beam patterns in a 20-element TMLA for MFRs.

Figure 7. The corresponding scanned beam patterns at the first sideband frequencies (|m| = 1) with ±30° shifting from the broadside direction and the unaltered fundamental pattern, obtained with the QNM-GA-based time scheme presented in Fig. 6.

Shaped flat-top beam patterns

To obtain the shaped flat-top pattern, uniformly excited (In = 1) 16-element TMLA with equally spaced (d = 0.5λ) elements is optimized by QNM-GA, and the optimal switching scheme and the obtained radiation beam patterns are given in Figs. 8 and 9, respectively. The synthesized shaped pattern at the first sideband shows a wideband flat-top beam pattern of more than 30° width. A sharp transition of 7.56° is achieved with SLLs of −20.62 dB and −20.32 dB for the fundamental pattern and first sideband pattern, respectively. The power dissipated in flat-top shaped beam is measured as 21.2770% of total dissipated power, whereas the fundamental pattern used 34.2112% of the radiated power. 76.7652% of radiated power is utilized for the desired synthesized patterns, and 23.2348% of the power is dissipated in higher-order harmonics. The optimized ON times and starting times are presented in Table 2, and the results are shown in Table 3.

Figure 8. Switching sequence obtained with QNM-GA to generate the flat-top beam pattern in a 16-element TMLA for shaped pattern synthesis.

Figure 9. The corresponding shaped beam patterns at the first sideband (m = 1), obtained with the QNM-GA-based time scheme shown in Fig. 8.

Cosecant-squared beam patterns

To obtain the cosecant-squared pattern, uniformly excited (In = 1) 20-element TMLA with equally spaced (d = 0.5λ) elements is optimized by QNM-GA, and the optimal switching scheme and the obtained radiation beam patterns are given in Figs. 10 and 11, respectively. The synthesized cosecant-squared beam pattern at the first sideband as well as the fundamental pattern has also been considered for SLL reduction. The SLLs of −20.17 dB and −18.6 dB is achieved for the fundamental pattern and cosecant-shaped first sideband pattern, respectively. The power dissipated in desired patterns is measured as 84.18% of total dissipated power, whereas the remaining 15.82% is wasted in higher sidebands. The optimized ON times and starting times are presented in Table 2, and the results are shown in Table 3.

Figure 10. Switching sequence obtained with QNM-GA to generate the cosecant-squared beam pattern in a 20-element TMLA for cosecant pattern synthesis.

Figure 11. The corresponding cosecant-squared beam pattern at the first sideband (m = 1), obtained with the QNM-GA-based time scheme shown in Fig. 10.

The optimized ON times and starting times are reported in Table 1. The results and the comparison with other published results for the monopulse and scanned beam patterns are presented in Tables 2 and 3, respectively. The SLLs achieved for all the multi-patterns are relatively smaller than the other reported works. It is observed that the SLL0 of monopulse and steered patterns are reduced to −40.89 dB and −33.12 dB compared to the best-reported results of −40 dB [Reference Yang, Chen and Nie9] and −21.7 dB [Reference Rocca, Zhu, Bekele, Yang and Massa35], respectively. The proposed method also shows improvement in terms of SLL1 as −22.54 dB and −23.58 dB is achieved compared to −20.1 dB [Reference Yang, Chen and Nie9] and −21.6 dB [Reference Rocca, Zhu, Bekele, Yang and Massa35] of the best-reported works. From the comparison, it is clear that multiple patterns can be efficiently achieved with the proposed method. The total radiated power used by the synthesized multi-beams and the wasted power in other harmonic frequencies for all the proposed applications are presented in Fig. 12.

Figure 12. Dissipated power in proposed multi-beams and unused power in higher harmonics obtained with QNM-GA-based switching schemes.

Conclusion

This work has proposed and verified an unorthodox but efficient substitute of traditional antenna arrays for multi-pattern synthesis. Multiple radiation patterns for monopulse and MFRs systems are investigated with fourth-dimensional (4D) antenna arrays, where the radiation patterns of the array are synthesized by controlling the ON times of each element. The monopulse scanning patterns are investigated by creating concurrent sum–difference beams with controlled radiating properties. The “fourth-dimensional” concept is further extended for MFRs where simultaneously scanned and shaped patterns are beneficial. Apart from the narrowband patterns for MFRs, the same concept is also exploited for wideband applications like shaped flat-top and cosecant-squared beam patterns. For all the desired beam patterns, utmost care has been taken to minimize the wanted inter-references. Furthermore, the higher-order harmonics inherently generated due to time-modulation are also suppressed to save power. All the examples are supported with simulated results and also compared with its counterpart results whenever applicable (and available) to show the enormous possibilities toward an alternate solution of phased antenna arrays. The “time-modulated” arrays can be implemented with high-speed switches eliminating costly phase shifters required in traditional phased arrays. The erroneous attenuators can also be discarded with uniform excitations. The dynamic range ratio obtained is 1, which is an added advantage of this work compared to traditional arrays. The SLLs of the monopulse beams are lowered below the ultra-low level (–40 dB). The SLLs of the scanned beams are also suppressed below −30 dB, whereas for wideband patterns, all the SLLs are suppressed below −20 dB. The power wasted in unwanted directions for all the patterns are also suppressed to improve the efficiencies. Thus, a cost-effective alternate to phased arrays for multi-pattern synthesis is proposed in this article, and the idea can be exploited further in near future.

Financial Support

This work is endorsed under the grant of EEQ/2021/000700 provided by SERB-DST.

Competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Avishek Chakraborty passed his B.Tech. degree in Electronics and Communication Engineering. He received his M.Tech. degree in Radio Physics and Electronics with a specialization in Space Science and Microwaves from the University of Calcutta, West Bengal, India, in 2017. He has completed his Ph.D. in Electronics and Communication Engineering from National Institute of Technology (NIT) Durgapur in 2021. He has also successfully completed a DST-SERB sponsored research project by serving as DST-SERB Research Fellow in his Ph.D. tenure. He is presently working as an Assistant Professor at the Department of Electrical, Electronics & Communication Engineering (EECE), GITAM University, Bangalore. He has authored several research papers published in peer-reviewed international journals, conferences, and book chapters. He has also served as a reviewer of IEEE Access, International Journal of RF and Computer-Aided Engineering, Circuits Systems and Signal Processing, The Journal of Supercomputing, etc. His current research interests include antenna array synthesis, phased arrays, antenna array optimization using soft computing and artificial intelligence, and signal processing for antenna arrays and radars.

Ravi Shankar Saxena received his B.E. degree in Electronics and Communication Engineering from Rohilkhand University, India, in 2000. He has completed his M.Tech. from MNNIT Allahabad in 2004, and D.Phil. in Communication Technology from Department of Electronics and Communication, University of Allahabad in 2016. He has published more than 16 research papers in different national and international journals and conference proceedings. His areas of interests include Patch antenna, Experimental nanoscience including Nanostructure fabrication, Optoelectronics properties, and Energy technologies. He is currently working as a Professor in GMRIT, Razam, Andhra Pradesh, India.

Anshoo Verma is working as Assistant Professor in Department of Civil Engineering, IES Institute of Technology and Management at IES University, Bhopal, since last 3 years. He has experience of more than 3 years in academics and research. Dr. Verma has completed his Master of Technology in Construction Technology and Management from Oriental Institute of Science & Technology in 2021.

Ashima Juyal has been working in the Department of Electronics & Communication Engineering at Uttaranchal Institute of Technology, Uttaranchal University, Dehradun, India. She has experience of more than 5 years in research and academics. Ms. Ashima is currently deputed at Research and Innovations Division at Uttaranchal University. Her areas of interests are Internet of Things, fuzzy logics, blockchain, and deep learning.

Sumit Gupta is working as Assistant Professor in the Department of Electronics and Communication at SR University, boasts over 13 years of experience in the field of educational profession and research, having previously served as a faculty member at KL University. He completed Ph.D. in Optical Multiplexing Technique in Fiber Communication from the prestigious National Institute of Technology, Bhopal, completed in 2017. He published his work in various reputed journals. Dr. Gupta currently focuses his research efforts on the cutting-edge fields of quantum communication and free space optical technology, exploring advanced multiplexing techniques.

Indrasen Singh is working as an Assistant Professor in the School of Electronics Engineering, VIT Vellore, Tamil Nadu, India. He received his B.Tech. and M.Tech. Degrees in electronics and communication engineering from Uttar Pradesh Technical University, Lucknow, India, in 2006, and 2010, respectively. He obtained his Ph.D. degree in electronics and communication engineering from National Institute of Technology Kurukshetra, Haryana, India, in 2019. He has more than 12 years of teaching, research experience in various reputed technical institutes or universities. He is the editorial board member of AJECE, Science Publishing Group, USA. He is the reviewer of many international journals and served as TPC member and reviewer in different conferences. He has published more than 20 research papers in national/international journals/conferences of repute and many are under review. His research interests are in the area of cooperative communication, stochastic geometry, modeling of wireless, heterogeneous networks, millimeter wave communications, device-to-device, and 5G/6G communication.

Gopi Ram completed his B.E. degree in Electronics and Telecommunication Engineering from Government Engineering College, Jagdalpur, Chhattisgarh, India, in the year 2007. He received his M.Tech. degree in Telecommunication Engineering from National Institute of Technology Durgapur, West Bengal, India, in the year 2011. He joined as a full-time Institute Research Scholar in the year of 2012 at National Institute of Technology, Durgapur, to carry out research for Ph.D. degree. He received the scholarship from the Ministry of Human Resource and Development (MHRD), Government of India, for the period 2009–2011 (M. Tech) and 2012–2015 (Ph.D). He is currently working as an Assistant Professor at the Department of Electronics and Communication Engineering, National Institute of Technology, Warangal, Telangana, India. His research interests include evolutionary optimization-based radiation pattern synthesis of antenna arrays, time-modulated antenna arrays, and applications of soft computing techniques in electromagnetics. He has published more than 50 research papers in international journals and conferences.

Durbadal Mandal completed his B.E. degree in Electronics and Communication Engineering from Regional Engineering College, Durgapur, West Bengal, India, in the year 1996. He received M.Tech. and Ph.D. degrees from National Institute of Technology, Durgapur, West Bengal, India, in the year 2008 and 2011, respectively. Presently, he is attached with National Institute of Technology Durgapur, West Bengal, India, as Associate Professor in the Department of Electronics and Communication Engineering. He has supervised 16 doctoral students and has completed 2 DST-SERB Sponsored Research Projects as Project Investigator till date. His research interest includes array antenna design; filter optimization via evolutionary computing techniques. He has published more than 350 research papers in international journals and conferences.

References

Hansen, RC (2009) Phased Array Antennas, 2nd edn. Chang Kai Wiley Series in Microwave and Optical Engineering. Hoboken, New Jersey: John Wiley & Sons.10.1002/9780470529188CrossRefGoogle Scholar
Mailloux, RJ (2006) A history of phased array antennas. In History of Wireless, Chang, Kai Wiley Series in Microwave and Optical Engineering. Hoboken, New Jersey: John Wiley & Sons, 567603.Google Scholar
Chakraborty, A, Ram, G and Mandal, D (2021) Time-domain approach towards smart antenna design. In Signals and Communication Technology, Wideband, Multiband, and Smart Antenna Systems Switzerland AG: Springer, Cham, 363394.Google Scholar
Rocca, P, Oliveri, G, Mailloux, RJ and Massa, A (2016) Unconventional phased array architectures and design methodologies—a review. Proceedings of the IEEE 104(3), 544560.10.1109/JPROC.2015.2512389CrossRefGoogle Scholar
Shanks, HE and Bickmore, RW (1959) Four-dimensional electromagnetic radiators. Canadian Journal of Physics 37, 263275.10.1139/p59-031CrossRefGoogle Scholar
Shanks, HE (1961) A new technique for electronic scanning. IRE Transactions on Antennas and Propagation 9, 162166.10.1109/TAP.1961.1144965CrossRefGoogle Scholar
Kummer, WH, Villeneuve, AT, Fong, TS and Terrio, FG (1963) Ultra-low sidelobes from time-modulated arrays. IEEE Transactions on Antennas and Propagation 11, 633639.10.1109/TAP.1963.1138102CrossRefGoogle Scholar
Haupt, RL (2017) Antenna arrays in the time domain: An introduction to timed arrays. IEEE Antennas and Propagation Magazine 59, 3341.10.1109/MAP.2017.2686082CrossRefGoogle Scholar
Yang, S, Chen, Y and Nie, Z (2008) Multiple patterns from time-modulated linear antenna arrays. Electromagnetics 28, 562571.10.1080/02726340802428671CrossRefGoogle Scholar
Chakraborty, A, Ram, G and Mandal, D (2022) Electronic beam steering in timed antenna array by controlling the harmonic patterns with optimally derived pulse-shifted switching sequence. In International Conference on Innovative Computing and Communications, 205216.10.1007/978-981-16-3071-2_18CrossRefGoogle Scholar
Rocca, P, Yang, F, Poli, L and Yang, S (2019) Time-modulated array antennas–theory, techniques, and applications. Journal of Electromagnetic Waves and Applications 33(12), 15031531.10.1080/09205071.2019.1627251CrossRefGoogle Scholar
Maneiro-Catoira, R, Brégains, J, García-Naya, JA and Castedo, L (2017) Time modulated arrays: From their origin to their utilization in wireless communication systems. Sensors 17(3), .10.3390/s17030590CrossRefGoogle ScholarPubMed
Fondevila, J, Brégains, JC, Ares, F and Moreno, E (2004) Optimizing uniformly excited linear arrays through time modulation. IEEE Antennas and Wireless Propagation Letters 3, 298301.10.1109/LAWP.2004.838833CrossRefGoogle Scholar
Yang, S, Gan, YB and Tan, PK (2003) A new technique for power-pattern synthesis in time-modulated linear arrays. IEEE Antennas and Wireless Propagation Letters 2, 285287.10.1109/LAWP.2003.821556CrossRefGoogle Scholar
Brégains, JC, Fondevila-Gómez, J, Franceschetti, G and Ares, F (2008) Signal radiation and power losses of time-modulated arrays. IEEE Transactions on Antennas and Propagation 56, 17991804.10.1109/TAP.2008.923345CrossRefGoogle Scholar
Yang, S, Gan, YB and Qing, A (2002) Sideband suppression in time-modulated linear arrays by the differential evolution algorithm. IEEE Antennas and Wireless Propagation Letters 1, 173175.10.1109/LAWP.2002.807789CrossRefGoogle Scholar
Yang, S, Can, YB and Tan, PK (2004) Comparative study of low sidelobe time modulated linear arrays with different time schemes. Journal of Electromagnetic Waves and Applications 18, 14431458.10.1163/1569393042954910CrossRefGoogle Scholar
Tennant, A and Chambers, B (2007) A two-element time-modulated array with direction-finding properties. IEEE Antennas and Wireless Propagation Letters 6, 6465.10.1109/LAWP.2007.891953CrossRefGoogle Scholar
Li, G, Yang, S, Chen, Y and Nie, Z (2009) A novel electronic beam steering technique in time modulated antenna arrays. Progress in Electromagnetics Research 97, 391405.10.2528/PIER09072602CrossRefGoogle Scholar
Tong, Y and Tennant, A (2010) Simultaneous control of sidelobe level and harmonic beam steering in time-modulated linear arrays. Electronics Letters 46, 200202.10.1049/el.2010.2629CrossRefGoogle Scholar
Poli, L, Rocca, P, Oliveri, G and Massa, A (2011) Harmonic beamforming in time-modulated linear arrays. IEEE Transactions on Antennas and Propagation 59, 25382545.10.1109/TAP.2011.2152323CrossRefGoogle Scholar
Tong, Y and Tennant, A (2012) A two-channel time modulated linear array with adaptive beamforming. IEEE Transactions on Antennas and Propagation 60, 141147.10.1109/TAP.2011.2167936CrossRefGoogle Scholar
Poli, L, Moriyama, T and Rocca, P (2014) Pulse splitting for harmonic beamforming in time-modulated linear arrays. International Journal of Antennas and Propagation 2014, 19.10.1155/2014/797590CrossRefGoogle Scholar
Chakraborty, A, Mandal, D and Ram, G (2019) Beam steering in a time switched antenna array with reduced side lobe level using evolutionary optimization technique. In 2019 IEEE Indian Conference on Antennas and Propagation, InCAP 2019.10.1109/InCAP47789.2019.9134497CrossRefGoogle Scholar
Ram, G, Panduro, MA, Reyna, A, Kar, R and Mandal, D (2018) Pattern synthesis and broad nulling optimization of STMLAA with EM simulation. International Journal of Numerical Modelling: Electronic Networks, Devices and Fields 31, .10.1002/jnm.2322CrossRefGoogle Scholar
Barton, DK (2010) History of monopulse radar in the US. IEEE Aerospace and Electronic Systems Magazine 25, c1c16.10.1109/MAES.2010.5464419CrossRefGoogle Scholar
Sherman, SM (1985) Monopulse principles and techniques. IEEE Antennas and Propagation Society Newsletter 27(5), .Google Scholar
Chakraborty, A, Ram, G and Mandal, D (2021) Pattern synthesis of timed antenna array with the exploitation and suppression of harmonic radiation. International Journal of Communication Systems 34, .10.1002/dac.4727CrossRefGoogle Scholar
Chakraborty, A, Ram, G and Mandal, D (2021) Time-modulated multibeam steered antenna array synthesis with optimally designed switching sequence. International Journal of Communication Systems 34, .10.1002/dac.4828CrossRefGoogle Scholar
Chakraborty, A, Ram, G and Mandal, D (2021) Multibeam steered pattern synthesis in time-modulated antenna array with controlled harmonic radiation. International Journal of RF and Microwave Computer-Aided Engineering 31, .10.1002/mmce.22597CrossRefGoogle Scholar
Chakraborty, A, Ram, G and Mandal, D (2020) Optimal pulse shifting in timed antenna array for simultaneous reduction of sidelobe and sideband level. IEEE Access 8, 131063131075.10.1109/ACCESS.2020.3010047CrossRefGoogle Scholar
Chakraborty, A, Ram, G and Mandal, D (2022) Time-modulated linear array synthesis with optimal time schemes for the simultaneous suppression of sidelobe and sidebands. International Journal of Microwave and Wireless Technologies 14, 768780.10.1017/S175907872100088XCrossRefGoogle Scholar
Bhattacharya, R, Saha, S and Bhattacharyya, TK (2017) Mutated IWO optimized 4-D array for femtocell cognitive radio. IEEE Antennas and Wireless Propagation Letters 16, 26142617.10.1109/LAWP.2017.2735999CrossRefGoogle Scholar
Chakraborty, A, Ram, G and Mandal, D (2019) Power pattern synthesis of a moving phase center time modulated antenna array using symmetrically and asymmetrically positioned time schemes. International Journal of RF and Microwave Computer-Aided Engineering 32(12), .Google Scholar
Rocca, P, Zhu, Q, Bekele, ET, Yang, S and Massa, A (2014) 4-D arrays as enabling technology for cognitive radio systems. IEEE Transactions on Antennas and Propagation 62, 11021116.10.1109/TAP.2013.2288109CrossRefGoogle Scholar
Poddar, S, Paul, P, Chakraborty, A, Ram, G and Mandal, D (2022) Design optimization of linear arrays and time‐modulated antenna arrays using meta‐heuristics approach. International Journal of Numerical Modelling: Electronic Networks, Devices and Fields 35(5), .10.1002/jnm.3010CrossRefGoogle Scholar
De Jong, K, Fogel, DB and Schwefel, H-P (2004) A history of evolutionary computation. In Baeck, Thomas, Fogel, DB and Michalewicz, Z (eds.), Handbook of Evolutionary Computation, 1st edn. Oxford, United Kingdom: Oxford University Press, .Google Scholar
Del Ser, J, Osaba, E, Molina, D, Yang, XS, Salcedo-Sanz, S, Camacho, D, Das, S, Suganthan, PN, Coello, CA and Herrera, F (2019) Bio-inspired computation: Where we stand and what’s next. Swarm and Evolutionary Computation 48, 220250.10.1016/j.swevo.2019.04.008CrossRefGoogle Scholar
Ram, G, Mandal, D, Kar, R and Ghoshal, SP (2014) Optimized hyper beamforming of receiving linear antenna arrays using Firefly algorithm. International Journal of Microwave and Wireless Technologies 6, 181194.10.1017/S175907871300086XCrossRefGoogle Scholar
Durmus, A and Kurban, R (2021) Optimum design of linear and circular antenna arrays using equilibrium optimization algorithm. International Journal of Microwave and Wireless Technologies 13, 986997.10.1017/S1759078720001774CrossRefGoogle Scholar
Rattan, M, Patterh, MS and Sohi, BS (2009) Optimization of circular antenna arrays of isotropic radiators using simulated annealing. International Journal of Microwave and Wireless Technologies 1, 441446.10.1017/S1759078709990687CrossRefGoogle Scholar
Das, A, Mandal, D and Kar, R (2021) An optimal circular antenna array design considering the mutual coupling employing ant lion optimization. International Journal of Microwave and Wireless Technologies 13, 164172.10.1017/S1759078720000914CrossRefGoogle Scholar
Dib, NI (2015) Synthesis of thinned planar antenna arrays using teaching-learning-based optimization. International Journal of Microwave and Wireless Technologies 7, 557563.10.1017/S1759078714000798CrossRefGoogle Scholar
Bogdan, G, Godziszewski, K and Yashchyshyn, Y (2020) Experimental investigation of beam-steering applied to 2 × 2 MIMO system with single receiving RF chain and time-modulated antenna array. International Journal of Microwave and Wireless Technologies 12, 504512.10.1017/S1759078720000744CrossRefGoogle Scholar
Dib, N and Sharaqa, A (2015) Design of non-uniform concentric circular antenna arrays with optimal sidelobe level reduction using biogeography-based optimization. International Journal of Microwave and Wireless Technologies 7, 161166.10.1017/S1759078714000610CrossRefGoogle Scholar
Ram, G, Mandal, D, Kar, R and Ghoshal, SP (2017) Directivity improvement and optimal far field pattern of time modulated concentric circular antenna array using hybrid evolutionary algorithms. International Journal of Microwave and Wireless Technologies 9, 177190.10.1017/S1759078715001075CrossRefGoogle Scholar
Misra, B and Mahanti, GK (2022) Meta-heuristic optimization algorithms for synthesis of reconfigurable hexagonal array antenna in two principle vertical planes. International Journal of Microwave and Wireless Technologies 14, 158165.10.1017/S1759078721000337CrossRefGoogle Scholar
Neri, F and Cotta, C (2012) Memetic algorithms and memetic computing optimization: A literature review. Swarm and Evolutionary Computation 2, 114.10.1016/j.swevo.2011.11.003CrossRefGoogle Scholar
Weile, DS and Michielssen, E (1997) Genetic algorithm optimization applied to electromagnetics: A review. IEEE Transactions on Antennas and Propagation 45, 343353.10.1109/8.558650CrossRefGoogle Scholar
Zhang, X, Lin, M, Zhang, X and Li, Y (2019) The design of microstrip array antenna and its optimization by a memetic method. IEEE Access 7, 9643496443.10.1109/ACCESS.2019.2929557CrossRefGoogle Scholar
Li, D, Qi, L and Roshchina, V (2008) A new class of quasi-Newton updating formulas. Optimization Methods & Software 23, 237249.10.1080/10556780701646360CrossRefGoogle Scholar
Zhou, W and Zhang, L (2020) A modified Broyden-like quasi-Newton method for nonlinear equations. Journal of Computational and Applied Mathematics 372, .10.1016/j.cam.2020.112744CrossRefGoogle Scholar
Fang, X, Ni, Q and Zeng, M (2018) A modified quasi-Newton method for nonlinear equations. Journal of Computational and Applied Mathematics 328, 4458.10.1016/j.cam.2017.06.024CrossRefGoogle Scholar
Krishna Chaitanya, R, Raju, GSN, Raju, KVSN and Mallikarjuna Rao, P (2022) Antenna pattern synthesis using the quasi Newton method, firefly and particle swarm optimization techniques. IETE Journal of Research 68(2), 11481156.10.1080/03772063.2019.1643263CrossRefGoogle Scholar
Morabito, AF, Di Carlo, A, Di Donato, L, Isernia, T and Sorbello, G (2019) Extending spectral factorization to array pattern synthesis including sparseness, mutual coupling, and mounting-platform effects. IEEE Transactions on Antennas and Propagation 67, 45484559.10.1109/TAP.2019.2905977CrossRefGoogle Scholar
Palmeri, R, Isernia, T and Morabito, AF (2019) Diagnosis of planar arrays through phaseless measurements and sparsity promotion. IEEE Antennas and Wireless Propagation Letters 18, 12731277.10.1109/LAWP.2019.2914529CrossRefGoogle Scholar
Ram, G (2021) Multi-beam steered harmonic pattern synthesis in timed antenna array with optimized and pre-defined RF switching. International Journal of Numerical Modelling: Electronic Networks, Devices and Fields 34, .10.1002/jnm.2912CrossRefGoogle Scholar
Figure 0

Figure 1. N-element array connected through switches and managed by a CPLD.

Figure 1

Figure 2. (a) Simple switching sequence and (b) shifted switching sequence for nth radiator in TMLA.

Figure 2

Figure 3. Flow chart of QNM-based memetic GA optimization.

Figure 3

Figure 4. Switching sequence obtained with QNM-GA to get the sum–difference beam pattern in a 20-element TMLA for MTR.

Figure 4

Table 1. Control parameters of QNM-based memetic GA optimization

Figure 5

Figure 5. The corresponding sum pattern at central frequency and the squinted difference pattern at first harmonic obtained with QNM-GA-based optimal in and out-of-phase time sequence presented in Fig. 4.

Figure 6

Table 2. Optimized ON-time period and starting times for multi-pattern synthesis

Figure 7

Table 3. Obtained simulation results for multi-pattern synthesis using QNM-GA

Figure 8

Table 4. Comparison of the proposed work for multiple pattern generation

Figure 9

Figure 6. Switching sequence obtained with QNM-GA to generate scanned beam patterns in a 20-element TMLA for MFRs.

Figure 10

Figure 7. The corresponding scanned beam patterns at the first sideband frequencies (|m| = 1) with ±30° shifting from the broadside direction and the unaltered fundamental pattern, obtained with the QNM-GA-based time scheme presented in Fig. 6.

Figure 11

Figure 8. Switching sequence obtained with QNM-GA to generate the flat-top beam pattern in a 16-element TMLA for shaped pattern synthesis.

Figure 12

Figure 9. The corresponding shaped beam patterns at the first sideband (m = 1), obtained with the QNM-GA-based time scheme shown in Fig. 8.

Figure 13

Figure 10. Switching sequence obtained with QNM-GA to generate the cosecant-squared beam pattern in a 20-element TMLA for cosecant pattern synthesis.

Figure 14

Figure 11. The corresponding cosecant-squared beam pattern at the first sideband (m = 1), obtained with the QNM-GA-based time scheme shown in Fig. 10.

Figure 15

Figure 12. Dissipated power in proposed multi-beams and unused power in higher harmonics obtained with QNM-GA-based switching schemes.