WLAN/WIMAX patch antenna design using artificial neural networks

doi:10.21203/rs.3.rs-1991392/v1

Download PDF

Research Article

WLAN/WIMAX patch antenna design using artificial neural networks

https://doi.org/10.21203/rs.3.rs-1991392/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

In this article, we propose a coplanar waveguide (CPW) antenna for WLAN/WIMAX applications using the artificial neural network (ANN) technique. In order to attain the correct resonant frequencies for WLAN/WIMAX applications, we inserted three slots in the radiating patch to form an inverted U-shaped slot. To achieve antenna performance in the standard WLAN/WIMAX frequency band, we adjusted the position and dimensions of the inverted U-shaped slot. The ANN method helped to better predict the dimensions of the inverted U-shaped slot. To collect the datasets used for training the ANN structure, we wrote a Visual Basic (VB) script in order to control the HFSS simulator using MATLAB. After training the network, the dimensions of the inverted U-shaped slot could easily be predicted for the desired resonant frequencies. The predicted ANN dimensions were used to design the antenna using both CST and HFSS electromagnetic simulators. To better evaluate the CPW antenna performance, we study and discuss the reflection coefficient, radiation pattern, voltage standing wave ratio (VSWR), current distribution, gain, and radiation efficiency. A prototype of the proposed antenna was fabricated; the obtained results show |S₁₁| ≤ − 10 dB in the frequency bands of 2.31–2.43 GHz, 3.10–3.44 GHz, and 3.94–5.88 GHz, which make the antenna suitable for WLAN/WIMAX applications.

Microstrip patch antenna

coplanar waveguide antenna

resonant frequency

artificial neural networks

WLAN/WIMAX

Because of their many attractive characteristics, such as, small size, light weight, and low fabrication cost, microstrip patch antennas have garnered the interest of many researchers and designers. Coplanar waveguide (CPW) antennas have attracted interest particularly due to their simple feed and ability to cover various frequency bands. The simple microstrip patch structure has a radiating patch of different geometries, such as rectangular, triangular, circular, or square. The radiating patch is etched on top of a dielectric substrate. The trend is to use a single antenna device of compact size for many services and applications, such as a wireless local area network (WLAN) and a worldwide interoperability for microwave access (WiMAX) network. For this reason, the design of a multiband antenna that is able to integrate these applications is highly demanded. The IEEE 802.11 standards specify a spectrum for WLAN and WIMAX, with WLAN centered at 2.4 (2.42.483), 5.2 (5.155.35), and 5.8 (5.7255.825), whereas the WIMAX system works at 2.5 (2.52.69), 3.5 (3.403.69), and 5.5 (5.255.85).

Recently, several antenna configurations have been designed and presented to achieve a multiband antenna satisfying the WLAN standards at the 2.4/5.2/5.8 GHz frequency bands, and the WIMAX recommendations for the 2.5/3.5/5.5 GHz bands [1–10]. Among the configurations, inverted U-shaped slots printed on the metal patch [1], L-shaped, inverted L, and U-shaped slots embedded in the radiator element [2–4], a compact F-shaped antenna [5], a wide-slot monopole antenna [6], a double L-slot patch antenna [7], a ring monopole antenna [8], and a multiband antenna were realized by introducing square-loop elements into the square patch [9], and a dual broadband WLAN antenna [10]. In [11], a novel tri-band square-slot antenna with two identical L-strips was presented for WLAN and WiMAX communication. The proposed antenna consisted of two L-strips, a square slot, and a monopole radiator. However, to attain the required frequencies exactly seemed to be difficult manually; for this reason, the use of a technique to better find the optimized dimensions for the desired resonant frequencies is recommended.

In [12], the genetic algorithm technique was used as an optimization tool to design a tri-band patch antenna, the technique helped to determine the optimized dimension and location of the loaded elements in the radiating patch. In other research, metamaterials were used to create a multiband antenna. In [13, 14], the design of a compact broadband antenna was achieved by using metamaterial. In [15], function in the WIMAX frequency band was achieved by inserting a rectangular split ring resonator in a radiating element patch. Single cell metamaterial loading was employed to cover the required bands [16]. However, the designed antennas displayed some disadvantages; they either presented a low bandwidth, so they could not cover all the WLAN/WIMAX frequency bands, or they were bulky in size.

ANNs have been widely utilized for microwave modeling and design due to the fastest evaluation of the ANN from input to output [17–22]. ANN techniques have also been proposed by many papers as an alternative tool in the design of microstrip patch antenna. In [23], a neural network was employed to design rectangular microstrip antennas. Artificial neural networks were used as a tool to study the parametric analysis of a slot-loaded microstrip line feed patch antenna in [24]. As reported in [25], an ANN application with multilayer perceptrons (MLPs) was proposed to determinate the resonant frequency of E-shaped compact microstrip antennas (ECMAs). In the work [26], an ANN model was presented to design a patch antenna with an X-shaped slot embedded in the radiating patch. The study of microstrip patch antennas with the ANN technique using analysis and synthesis methods has been reported in [27]. In [28], by predicting the slot size using ANNs, the performance of rectangular patch antennas was improved. Yet in these articles, the ANN technique was mainly used to design antennas for mono-band applications, and the parameters of the proposed antennas were not presented clearly.

The purpose of this work is to design a CPW patch antenna by inserting three slots joined to form an inverted U-shaped slot in the metal patch. To generate multiple resonant frequencies for WAN/WIMAX applications, an ANN model with MLPs is proposed to predict the slot size over the required resonant frequencies. The training and testing datasets are generated by MATLAB using a script to help us to control the HFSS simulator, and the developed code allows us to collect data by saving the size of the inverted U-shaped slot and the resonant frequencies at each iteration. All datasets are organized into two matrices to feed the ANN network as the second step of this study.

The configuration of the proposed antenna is presented in Fig. 1. As depicted, the antenna is printed on a dielectric substrate of FR4 epoxy, a material that is inexpensive and commercially available, and in particular, offers a good compromise between performance and antenna design. This material is characterized by its dimensions (L × W × h), a relative permittivity of ε_r = 4.3, and a loss tangent of tanδ = 0.002. A 50-ohm feed line of width w_f and length l_f was used as a CPW transmission line to feed the patch antenna. On each side of the CPW feed line, two symmetrical finite ground planes are printed with the same dimensions (l_g × w_g). The feed line and the coplanar ground plane are separated by a gap (g). At the limit of the feed line, a rectangular patch as the main element of the antenna structure is etched on the substrate with dimensions (l_p × w_p). The patch and the ground plane consist of copper with a thickness of t = 0.035 mm. An inverted U-shaped slot is embedded in the radiating patch, as shown in Fig. 1. The overall dimensions and parameters of the antenna components are listed in Table 1.

Table 1

Antenna parameters.
Component	Parameters
Substrate	L = 33 mm
(FR4 epoxy (4.3))	W = 24 mm h = 1.56 mm
Ground plane	l_g = 10.5 mm
(Copper)	w_g = 9.9 mm g = 0.5 mm thickness = 0.035 mm
Patch	l_p = 17 mm
(Copper)	w_p = 16 mm thickness = 0.035 mm
Feed line	l_f = 13.5 mm
(Copper)	w_f = 3.2 mm thickness = 0.035 mm

ANNs mimic the human brain, hence they are described as artificial. An ANN can be considered to be a highly simplified model inspired by the structure and performance of a biological neural network [29, 30], and recently has become an alternative modeling and design tool for numerical and analytical modeling [31]. In terms of their structure, we can find two types of neural networks: feedforward networks and recurrent networks [32]. Recurrent networks are networks where the outputs of some neurons are fed back to the same neurons or to neurons in previous layers. Recurrent networks are said to have a dynamic memory [32]. MLPs are found in the common feedforward ANN model where signals flow in a single direction from the input to the output of the network.

To build a neural model, the first step is the generation and collection of datasets for training and testing the model. These datasets can be either measured, calculated, or simulated. For this reason, we have written a VB script so as to control the HFSS electromagnetic simulator in MATLAB. The developed code varies the dimensions (X₁, Y₁, X₂, Y₂) of the inverted U slot embedded in the CPW patch antenna (Fig. 2) over a specified range, as given in Table 2, and records the resonant frequencies for each iteration while the position (x₁, y₁) is kept constant. Hence, two data matrices are achieved [D] = [X₁ Y₁ X₂ Y₂] and [F] = [f₁ f₂ f₃]. A total of 1600 pieces of data are generated, of which 1280 samples are used for training, and the rest are used for testing. A sample of data is given in Table 3. For the presented ANN model, we used a reverse synthesis. The datasets are arranged in matrices, where the input vectors consist of the resonant frequencies while the output vectors consist of the dimensions of the slots.

Table 2

ANN datasets.
Slot size	Specified range	Step size	Generated samples
X₁	0.55 ≤ X₁ ≤ 1 mm	0.05 mm	1600 samples 1280 for training 320 for testing
Y₁	12 ≤ Y₁ ≤ 13.5 mm	0.5 mm
X₂	11.5 ≤ X₂ ≤ 13 mm	0.5 mm
Y₂	0.45 ≤ Y₂ ≤ 0.9 mm	0.05 mm

As seen in Fig. 2, the three slots (slot UP, slot L, and slot R) are unified in order to form an inverted U-shaped slot. The position and the dimension of the embedded slots are illustrated in Fig. 2. By adjusting the dimension of the slots and their positions, and to obtain better optimization, one can improve various antenna parameters such as the voltage standing wave ratio (VSWR), reflection coefficient, and gain.

To train the datasets, we used MLPs, which is the most commonly used and efficient model of artificial neural networks with a number of layers. The creation of an MLP network structure requires the determination of the appropriate training algorithm, the number of layers, and the number of neurons in each layer.

To determine the appropriate training algorithm for the MLP model, eight different algorithms were used for training the datasets of the proposed antenna configuration: resilient backpropagation (RP), Polak-Ribiere conjugate gradient (CGP), one step secant (OSS), scaled conjugate gradient (SCG), Powell-Beale conjugate gradient (CGB), Bayesian regulation (BR), conjugate gradient with FletcherPeeves (CGF), and Levenberg-Marquardt (LM).

Table 3

Sample of datasets.
X₁	Y₁	X₂	Y₂	f₁	f₂	f₃
0.55	12	11.5	0.45	2.621	2.9855	3.953
0.55	12	11.5	0.65	2.6705	3.404	5.4965
0.6	12.5	11.5	0.6	2.6795	3.647	5.4515
0.6	12.5	11.5	0.9	2.549	3.548	5.447
0.7	12	12.5	0.9	2.4995	2.9855	3.926
0.7	12.5	11.5	0.45	2.747	3.638	5.474
0.8	12.5	13	0.7	2.45	3.2375	3.5975
0.8	13	11.5	0.55	2.558	2.9675	4.034
0.8	13	12	0.6	2.5985	3.647	5.3975
0.8	13	12	0.65	2.5265	3.1835	3.665
A suitable training algorithm for the proposed ANN structure was selected by evaluating statistical criteria for different training algorithms. The statistical criteria used were root mean square error (RMSE), mean absolute error (MAE), and mean absolute percentage error (MAPE). These statistical criteria are given by the following formulas:

Root mean Square error:

$$\text{R}\text{M}\text{S}\text{E}=\sqrt{\left(\frac{1}{\text{n}}\right)\sum _{\text{i}=1}^{\text{n}}{\left({\text{t}}_{\text{i}}-{{\alpha }}_{\text{i}}\right)}^{2}}$$

Mean absolute error:

$$\text{M}\text{A}\text{E}=\left(\frac{1}{\text{n}}\sum _{\text{i}=1}^{\text{n}}\left|{\text{t}}_{\text{i}}-{{\alpha }}_{\text{i}}\right|\right)$$

Mean absolute percentage error:

$$\text{M}\text{A}\text{P}\text{E}=\left(\frac{100}{\text{n}}\sum _{\text{i}=1}^{\text{n}}\left|\frac{{\text{t}}_{\text{i}}-{{\alpha }}_{\text{i}}}{{\text{t}}_{\text{i}}}\right|\right)$$

Where n is the total number of samples, and t_i and α_i represent, respectively, the target and output data. The statistical criteria for each training algorithm used to evaluate the performance of the built ANN structure are summarized in Table 4.

After several trials, by changing the number of hidden layers as well as the number of neurons in each hidden layer and observing the ANN structure performance we concluded that the proposed ANN structure, as depicted in Fig. 3, consists of 5 layers: the input layer with 3 neurons, 3 hidden layers with 35 neurons in each layer, and an output layer with 4 neurons. The training algorithm chosen was RP, which presented the minimum error between the desired and the output data (Table 4).

In order to evaluate the performance of the proposed ANN model, 1280 datasets were used during the training process while the remaining 320 samples were selected arbitrary and used during the testing process. The MAPE was calculated to be 8.69% for training data, whereas the obtained MAPE during the testing process is approximately 9.29%. These results confirm the validity of the proposed ANN model.

Table 4

Statistical criteria for different training algorithms.
Training algorithm	RMSE	MAE	MAPE
RP	0.281	0.199	8.816
CGP	0.285	0.203	8.871
OSS	0.292	0.210	9.346
SCG	0.286	0.204	8.970
CGB	0.286	0.205	9.023
BR	0.474	0.167	7.700
CGF	0.298	0.216	9.553
LM	0.925	0.265	9.005

The dimensions of the inverted U-shaped slot for the desired operating frequencies 2.5/3.5/5.5 GHz predicted by the ANN structure are depicted in Table 5:

Table 5

ANN predicted values.
Dimensions	ANN values
X₁	0.60 mm
Y₁	13.19 mm
X₂	11.69 mm
Y₂	0.70 mm

The geometric evolution of the presented antenna is illustrated in Fig. 4. From Fig. 5 (curve ANT-1) we can observe that, for the first antenna geometry Fig. 4(a), which starts with a conventional rectangular patch and a coplanar ground plane, two resonant frequencies are obtained at approximately 3.46 and 5.49 GHz, with a reflection coefficient of -24 dB and − 27 dB respectively. By including an inverted U-shaped slot on the radiating patch, as illustrated in antenna2 Fig. 4(b), and by predicting its optimized dimensions using the ANN technique (Table 5), we can observe three resonant modes and two distinct wide bands centered at approximately 2.51, 3.51, and 5.50 GHz, with a good reflection coefficient almost equal to -38 dB, -32 dB, and − 32.5 dB, respectively, as seen in Fig. 5 (curve ANT-2).

Therefore, the proposed antenna covers all the 2.4/5.2/5.8 GHz WLAN and 2.5/3.5/5.5 GHz WIMAX bands.

For better comprehension of antenna performance, various antenna parameters needed to be determined. For this reason, we studied and discussed the reflection coefficient, the radiation pattern, the VSWR, the surface current distribution, the gain, and the radiation efficiency of the proposed antenna, which was simulated based on the initial antenna dimensions presented in Table 1 and the predicted ANN values of the dimensions [X₁, Y₁, X₂, Y₂] given in Table 5 using both CST and HFSS electromagnetic simulators.

From the − 10 dB bandwidth simulated results illustrated in Fig. 6, it is clear that the antenna presents three resonant frequencies and two wide operating bands. The lower band has an impedance bandwidth of 140 MHz (2.40–2.58 GHz) centered at 2.51 GHz, while the impedance bandwidth for the upper band is 3260 MHz (3.00-6.41 GHz), which presents two resonant frequencies of 3.51 GHz and 5. 50 GHz. Using a Rohde & Schwarz ZVB20 vector network analyzer, we measured the reflection coefficient of the fabricated prototype antenna (Fig. 7), and the measured results showed that the antenna has |S₁₁| ≤ − 10 dB in the frequency bands of 2.31–2.43 GHz, 3.10–3.44 GHz, and 3.94–5.88 GHz (Fig. 6). These results prove that the proposed antenna can operate over 2.4/5.2/5.8 GHz WLAN and 2.5/3.5/5.5 GHz WIMAX bands. The measured and simulated S₁₁ are not exactly the same, this can be due to the fabrication process, manual soldering of SMA connector, and measurement circumstances.

To verify the radiation pattern characteristics, we measured the radiation pattern using the Geozondas principle as shown in Fig .8. In Fig. 9(a), Fig. 9(b), and Fig. 9(c) the measured and simulated far-field radiation patterns of the proposed antenna in the elevation xoz plane (E-plane) and the azimuthal xoy plane (H-plane) for the desired operating frequencies of 2.50 GHz, 3.50 GHz, and 5.50 GHz, are given respectively. From these figures, we note that over the three operating frequencies, the antenna presents an omnidirectional radiation pattern in the xoy plane, while it has a bidirectional pattern in the xoz plane. An excellent agreement between measured and simulated results is also noted.

To evaluate the impedance matching between the transmission line and the antenna, we used the VSWR parameter, which is a function of the reflection coefficient. A good antenna is defined by an acceptable value of VSWR (1 ≤ VSWR ≤ 2). The VSWR of the designed CPW antenna was simulated and plotted in Fig. 10. The VSWR of the proposed antenna is approximately 1.20, 1.04, and 1.05 at the resonant frequencies 2.50 GHz, 3.50 GHz, and 5.50 GHz, respectively, which indicates good matching.

In order to demonstrate the contribution of the embedded slots in the radiating element patch to the resonant frequencies, we present in Fig. 11 the proposed antennas simulated surface current distribution at 2.50, 3.50, and 5.50 GHz. From Fig. 11(a), it is clear that most of the surface currents are concentrated around the inverted U-shaped slot. This result indicates that the embedded slot is the essential contributor to the occurrence of the lower resonant mode at 2.50 GHz. It is observed from Fig. 11(b) and Fig. 11(c), that the surface current distribution is slightly around the inner and the outside of the slot, which leads with the help of the ANN technique to match perfectly the second and third resonant mode. Thus, from the surface current distribution, we can conclude that the embedded slot is important in creating the proposed antenna resonance and generating two wide bands covering the 2.4/5.2/5.8 GHz WLAN and 2.5/3.5/5.5 GHz WiMAX bands.

The antenna gain and the radiation efficiency across the operating frequency bands were also verified and plotted. As shown in Fig. 12, for the first operating band (2.4–2.69 GHz), the antenna has a peak value of 1.35 dBi at 2.47 GHz, and a maximum gain of 2.15 dBi in the middle band (3.403.69 GHz) at 3.69 GHz, whereas a gain of 3.17 dBi at 5.85 GHz is obtained in the upper operating band (5.15–5.85 GHz). The radiation efficiency varies approximately from 70–87%.

All the simulated results of the presented CPW fed microstrip patch antenna are summarized in Table 6:

It can be seen from Table 6 that for the presented antenna, fundamental parameters like the reflection coefficient, VSWR, gain, and radiation efficiency show reasonable characteristics over the 2.4/5.2/5.8 GHz WLAN bands and 2.5/3.5/5.5 GHz WIMAX bands. The antenna is perfectly matched with the transmission line, because the VSWR is between 1 and 2. These results also show that the proposed CPW antenna exhibits an acceptable gain value and radiation efficiency, which indicates a minimum loss within the structure of the antenna and at the input terminals. Hence, this antenna is recommended for the WLAN/WIMAX communication systems.

Table 6

Simulated results of the designed antenna.
Antenna performance		WLAN			WIMAX
Antenna performance	2.4	5.2	5.8	2.5	3.5	5.5
Reflection coefficient (dB)	-10	-22.65	-20.21	-21.87	-31.98	-32.43
VSWR	1.93	1.15	1.22	1.20	1.04	1.05
Gain (dBi)	1.23	2.78	3.13	1.01	1.85	3.00
Radiation efficiency (%)	74.21	86.99	85.88	69.07	82.78	86.79

To support the simulated results, we present in Table 7a comparison between the obtained results and the results found in the literature.

Table 7

Comparison of alternative WLAN/WIMAX antennas.
References	Operating bands [GHz]	Gain [dBi]	Size [mm × mm ]
[1]	2.402.52, 3.403.60, 5.06.0	1.62, 1.47, 2.96	10 × 26
[2]	2.392.69, 3.43.85, 4.557.85	2, 3, 2.5	15 × 15
[5]	2.0-2.76, 3.04-4.0, 5.2-6.0	1.5, 1.7, 3.05	19 × 25
[12]	2.25-4.00	1.5, 2	14.8 × 29.6
[33]	2.40 -2.4835, 3.40–3.61	7, 7.4	60 × 45
[34]	2.36–2.70, 3.35–3.74, 5.01–6.12	0.94, 1.45, 2.61	32 × 12
[35]	2.20–2.52, 3.30–4.20	1.4, 1.6	14.75 × 26
Proposed	2.40–2.58, 3.00-6.41	1.01, 1.85, 3.00	24 × 33

Considering the covered bandwidths, gain, and antenna size, we conclude that compared with the references listed in Table 7, the designed CPW antenna has a simpler structure that can be fabricated easily, and can cover all the frequency bands of WLAN/WIMAX applications.

In this paper, a CPW patch antenna fabricated on a FR4 epoxy substrate of dimensions 24 × 33 × 1.56 mm³ was presented. The antenna was designed with the help of the ANN technique using MLPs. To generate multiple resonant frequencies for WLAN/WIMAX applications, we inserted an inverted U-shaped slot in the radiating patch, where the dimensions of the slot were predicted by the ANN. The simulated and measured results confirm that the ANN was able to predict with high accuracy the designed antenna parameters for the required operating frequencies. In terms of antenna performance, the proposed antenna covers all the operating bands of WLAN/WIMAX, with good radiation characteristics and acceptable gain. These features, in addition to its simple structure, make the proposed antenna a good candidate for WLAN/WIMAX systems.

I declare that the authors have no competing interests as defined by Springer, or other interests that might be perceived to influence the results and/or discussion reported in this paper.

CONFLICT OF INTEREST

All authors have seen and agree with the contents of the manuscript and there is no financial interest to report.

DATA AVAILABILITY STATEMENT

Not Applicable.

Kunwar, A., Gautam, A. K., & Rambabu, K. (2017). Design of a compact U-shaped slot triple band antenna for WLAN/WiMAX applications. AEU-International Journal of Electronics and Communications, 71, 82-88.
Moosazadeh, M., & Kharkovsky, S. (2014). Compact and small planar monopole antenna with symmetrical L-and U-shaped slots for WLAN/WiMAX applications. IEEE antennas and wireless propagation letters, 13, 388-391.
Chen, H., Yang, X., Yin, Y. Z., Fan, S. T., & Wu, J. J. (2013). Triband planar monopole antenna with compact radiator for WLAN/WiMAX applications. IEEE Antennas and Wireless Propagation Letters, 12, 1440-1443.
Kunwar, A., Gautam, A. K., & Kanaujia, B. K. (2017). Inverted L-slot triple-band antenna with defected ground structure for WLAN and WiMAX applications. International Journal of Microwave and Wireless Technologies, 9(1), 191-196.
Gautam, A. K., Kumar, L., Kanaujia, B. K., & Rambabu, K. (2015). Design of compact F-shaped slot triple-band antenna for WLAN/WiMAX applications. IEEE Transactions on Antennas and Propagation, 64(3), 1101-1105.
Liu, G., Liu, Y., & Gong, S. (2016). Compact tri‐band wide‐slot monopole antenna with dual‐ring resonator for WLAN/WiMAX applications. Microwave and Optical Technology Letters, 58(5), 1097-1101.
Chitra, R. J., & Nagarajan, V. (2013). Double L-slot microstrip patch antenna array for WiMAX and WLAN applications. Computers & Electrical Engineering, 39(3), 1026-1041.
Jo, S., Choi, H., Shin, B., Oh, S., & Lee, J. (2014). A CPW-fed rectangular ring monopole antenna for WLAN applications. International Journal of Antennas and Propagation, 2014.
Malik, J. S., Rafique, U., Ali, S. A., & Khan, M. A. (2017). Novel patch antenna for multiband cellular, WiMAX, and WLAN applications. Turkish Journal of Electrical Engineering and Computer Sciences, 25(3), 2005-2014.
Palandöken, M. (2017). Dual broadband antenna with compact double ring radiators for IEEE 802.11 ac/b/g/n WLAN communication applications. Turkish Journal of Electrical Engineering and Computer Sciences, 25(2), 1326-1333.
Hu, W., Yin, Y. Z., Fei, P., & Yang, X. (2011). Compact triband square-slot antenna with symmetrical L-strips for WLAN/WiMAX applications. IEEE Antennas and Wireless Propagation Letters, 10, 462-465.
Sami, G., Mohanna, M., & Rabeh, M. L. (2013). Tri-band microstrip antenna design for wireless communication applications. NRIAG Journal of Astronomy and Geophysics, 2(1), 39-44.
Zhou, C., Wang, G., Liang, J., Wang, Y., & Zong, B. (2014). Broadband antenna employing simplified MTLs for WLAN/WiMAX applications. IEEE antennas and wireless propagation letters, 13, 595-598.
Rajkumar, R., & Kiran, K. U. (2016). A compact metamaterial multiband antenna for WLAN/WiMAX/ITU band applications. AEU-International Journal of Electronics and Communications, 70(5), 599-604.
Rajeshkumar, V., & Raghavan, S. (2015). A compact metamaterial inspired triple band antenna for reconfigurable WLAN/WiMAX applications. AEU-International Journal of Electronics and Communications, 69(1), 274-280.
Huang, H., Liu, Y., Zhang, S., & Gong, S. (2014). Multiband metamaterial-loaded monopole antenna for WLAN/WiMAX applications. IEEE antennas and wireless propagation letters, 14, 662-665.
Zhang, Q. J., & Gupta, K. C. (2000). Neural networks for RF and microwave design. Artech House. Inc. Norwood, MA, USA.
Devabhaktuni, V. K., Yagoub, M. C., Fang, Y., Xu, J., & Zhang, Q. J. (2001). Neural networks for microwave modeling: Model development issues and nonlinear modeling techniques. International Journal of RF and Microwave Computer‐Aided Engineering, 11(1), 4-21.
Zhang, Q. J., Gupta, K. C., & Devabhaktuni, V. K. (2003). Artificial neural networks for RF and microwave design-from theory to practice. IEEE transactions on microwave theory and techniques, 51(4), 1339-1350.
Delgado, H. J., Thursby, M. H., & Ham, F. M. (2005). A novel neural network for the synthesis of antennas and microwave devices. IEEE Transactions on neural networks, 16(6), 1590-1600.
Kabir, H., Wang, Y., Yu, M., & Zhang, Q. J. (2008). Neural network inverse modeling and applications to microwave filter design. IEEE Transactions on Microwave Theory and Techniques, 56(4), 867-879.
Kabir, H., Zhang, L., Yu, M., Aaen, P. H., Wood, J., & Zhang, Q. J. (2010). Smart modeling of microwave devices. IEEE Microwave Magazine, 11(3), 105-118.
Türker, N., Güneş, F., & Yildirim, T. (2006). Artificial neural design of microstrip antennas. Turkish Journal of Electrical Engineering and Computer Sciences, 14(3), 445-453.
Aneesh, M., Ansari, J., & Singh, A. (2014). Analysis of microstrip line feed slot loaded patch antenna using artificial neural network. Progress In Electromagnetics Research B, 58, 35-46.
Akdagli, A., Toktas, A., Kayabasi, A., & Develi, I. (2013). An application of artificial neural network to compute the resonant frequency of E-shaped compact microstrip antennas. Journal of Electrical Engineering, 64(5), 317.
Aneesh, M., Singh, A., & Ansari, J. (2014). Investigations for performance improvement of X-shaped RMSA using artificial neural network by predicting slot size. Progress In Electromagnetics Research C, 47, 55-63.
Thakare, V. V., & Singhal, P. (2010). Microstrip antenna design using artificial neural networks. International Journal of RF and Microwave Computer‐Aided Engineering, 20(1), 76-86.
Khan, T., De, A., & Uddin, M. (2013). Prediction of slot-size and inserted air-gap for improving the performance of rectangular microstrip antennas using artificial neural networks. IEEE Antennas and Wireless Propagation Letters, 12, 1367-1371.
Simon, H. (1999). Neural networks: a comprehensive foundation. Prentice hall.
Yegnanarayana, B. (2009). Artificial neural networks. PHI Learning Pvt. Ltd..
Merad, L., Bendimerad, F. T., Meriah, S. M., & Djennas, S. A. (2007). Neural networks for synthesis and optimization of antenna arrays. RADIOENGINEERING-PRAGUE-, 16(1), 23.
Wang, J., & Kusiak, A. (Eds.). (2000). Computational intelligence in manufacturing handbook. CRC Press.
Liu, S., Wu, W., & Fang, D. G. (2015). Single-feed dual-layer dual-band E-shaped and U-slot patch antenna for wireless communication application. IEEE Antennas and Wireless Propagation Letters, 15, 468-471.
Kang, L., Wang, H., Wang, X. H., & Shi, X. (2014). Compact ACS‐fed monopole antenna with rectangular SRRs for tri‐band operation. Electronics letters, 50(16), 1112-1114.
Naidu, P. V. (2017). Printed V-shape ACS-fed compact dual band antenna for bluetooth, LTE and WLAN/WiMAX applications. Microsystem Technologies, 23(4), 1005-1015.

No competing interests reported.

Download PDF

Editor assigned by journal
06 Jul, 2023
Submission checks completed at journal
26 Aug, 2022
First submitted to journal
23 Aug, 2022

You are reading this latest preprint version

WLAN/WIMAX patch antenna design using artificial neural networks

Status:

Version 1

Abstract

Figures

1. Introduction

2. Proposed Antenna

3. Proposed Method

4. Results And Discussion

5. Conclusions

Declarations

References

Additional Declarations

Status:

Version 1