Single-Fed Broadband CP Bidirectional Antenna with Double-Layer Diagonally Aligned Plates for Universal UHF-RFID Applications

Research Center of Innovation Digital and Electromagnetic Technology (iDEMT), College of Industrial Technology, King Mongkut’s University of Technology North Bangkok, Bangkok, %ailand Faculty of Engineering, King Mongkut’s Institute of Technology Ladkrabang, Bangkok, %ailand Faculty of Engineering, Rajamangala University of Technology Rattanakosin, Nakhon Pathom, %ailand Department of Electrical and Computer Engineering, Faculty of Engineering, King Mongkut’s University of Technology North Bangkok, Bangkok, %ailand


Introduction
Radio frequency identification (RFID) technology is commonly utilized for wireless multiaccess identification and tracking in logistics, livestock farming, and toll collection [1,2]. Ultrahigh frequency (UHF) RFID varies from country to country. e RFID frequency band in China is 840.5 -844. 5 MHz, 920 -925 MHz in ailand, and 902 -928 MHz in the US. Meanwhile, the Federal Communications Commission's (FCC) universal RFID frequency band is 860 -960 MHz. RFID reader antennas have to be able to communicate between readers and tags, independent of tag orientation. To satisfy this requirement, CP antennas are thus necessary [3,4]. Meanwhile, bidirectional radiation is ideal for identification and tracking moving objects in a tunnel or on a conveyor belt [5].
is research proposes a single-fed broadband CP bidirectional antenna that efficiently covers the universal UHF-RFID frequency band. e proposed folded antenna is comprised of upper-layer conductor, lower-layer conductor, and wall patches. e upper-layer conductor consists of two diagonally aligned rectangular copper plates with a feeding gap at the center, and the lower-layer conductor is of two diagonally adjoined rectangular plates.
e upper-and lower-layer conductors are adjoined with the wall patches. e organization of the research is as follows: Section 1 is the introduction. Section 2 details the design of the antenna structure. Section 3 deals with the simulation and measurement results, and Section 4 discusses the parametric study analysis. e concluding remarks are provided in Section 5. Figure 1 illustrates the geometry and parameters of the single-fed broadband CP bidirectional antenna for universal UHF-RFID applications. e proposed folded antenna comprises upper-layer conductor, lower-layer conductor, and wall patches. e upper-layer conductor consists of two diagonally aligned rectangular plates with a feeding gap at the center, and the lower-layer conductor is of two diagonally adjoined rectangular plates. e upper-and lower-layer conductors are adjoined with the wall patches. Air is used as the substrate to achieve high gain, broad bandwidth, and low cost. e proposed broadband CP bidirectional antenna was fabricated from copper sheet of 1 mm in thickness. e selected copper thickness (1 mm) helped strengthen the antenna structure. e dimension of the proposed antenna was 146 mm × 275 mm × 10 mm (W × L × h). e antenna was excited by a 50 Ω coaxial feed. e width (W tf ) and length (L ty ) of the upper-layer rectangular patch were 74 and 138 mm, and the width (W bf ) and length (L by ) of the lowerlayer rectangular patch were 74 and 137.5 mm. e upperand lower-layer rectangular patches were staggered by 1 mm. e feeding gap (δ) was 3 mm fed by a coaxial cable. An N-type connector was connected to the coaxial cable hidden beneath the upper-layer plate. Table 1 tabulates the antenna parameters and optimal dimensions of the proposed broadband CP bidirectional antenna simulated by CST Microwave Studio [15]. Figure 2 shows three evolutionary stages of the proposed antenna: two diagonally aligned rectangular plates (upper-layer plates), the diagonally aligned rectangular plates with wall patches, and broadband CP bidirectional antenna. e initial length of the proposed antenna (L) was approximately λ/4 in order to achieve AR < 3 dB, where λ is the wavelength of the center frequency (900 MHz). e optimal length of the antenna was 275 mm. e wall patch dimension and feeding gap were optimized (Table 1). Circular polarization was thus realized as current flowed from the upper-layer plates to the wall patches and to the lower-layer plates. e time (t) for the current to flow from one end to the other end of the wall patch was different, with 90°phase difference.

Evolutionary Stages of the Antenna.
Figures 3(a)-3(c) show the simulated |S 11 |, impedance, and AR pattern of the proposed antenna in the three evolutionary stages. e double-layered structure of the proposed antenna resulted in AR bandwidth lower than 3 dB for the universal UHF-RFID frequency band. In Figure 3, Stage 1 of the antenna evolution achieved AR bandwidth close to 3 dB, and resistance and reactance were reduced by using the wall patches in Stage 2. In Stage 3, the resistance, reactance, and AR bandwidth were further reduced by incorporating the lower-layer rectangular plates.

Width of the Proposed Antenna (W). Figures 4(a) and 4(b)
show the simulated |S 11 | and AR of the proposed antenna under different antenna widths (W): 140, 143, 146, 149, and 152 mm. e simulation results revealed that the variation in W had minimal impact on |S 11 | but significant effect on AR bandwidth, especially in higher-frequency range. e AR bandwidth became narrower as W increased.
Specifically, as W decreased, the impedance bandwidth (|S 11 | < −10 dB) became wider, and AR amplitude deteriorated. On the other hand, as W increased, the impedance bandwidth shifted to lower frequency and became narrower, and AR amplitude deteriorated. e AR bandwidth nonetheless covered the universal UHF-RFID frequency band, independent of W.

2
International Journal of Antennas and Propagation e optimal width (W) of the proposed antenna was 146 mm, given that the AR bandwidth effectively covered the entire UHF-RFID frequency band. Besides, variations in W had no effect on the antenna gain in the lower frequency range and minimal effect in the upper frequency range. As a result, no graphical presentation of the effect of W on the antenna gain was provided. Meanwhile, as L decreased, the impedance bandwidth (|S 11 | < −10 dB) shifted to the higher-frequency range with slightly improved antenna gain. As L increased, the impedance bandwidth deteriorated with slightly lower antenna gain. Since variations in L had negligible effect on the antenna gain, no graphical presentation of the effect of L on the antenna gain was provided. Given the optimal wall height of 10 mm (Table 1), as h increased (12 and 14 mm), the AR bandwidth of the antenna improved and became narrower. As h decreased (6 and 8 mm), the AR bandwidth became wider. In addition, as h increased (12 mm), the impedance bandwidth (|S 11 | < −10 dB) slightly shifted at the same frequency range with steady antenna gain. As h varied (6, 8, and 14 mm), the impedance bandwidth deteriorated with slightly lower antenna gain. Since variations in h had negligible effect on the antenna gain, no graphical presentation of the effect of h on the antenna gain was provided.

Length of Feeding Gap (δ).
Given the optimal feeding gap (δ) at the center of the upper-layer conductor of 3 mm (Table 1), the simulated impedance bandwidth (|S 11 | < −10 dB) covered the universal UHF-RFID frequency band, as shown in Figure 7(a). Figures 7(a) and 7(b) illustrate the simulated |S 11 | and AR of the proposed antenna under variable feeding gaps (δ): 1, 3, 5, 7, and 9 mm. In Figure 7(a), as δ increased (5, 7, and 9 mm), the impedance bandwidth shifted to the lower frequency range, and the impedance bandwidth shifted to the higherfrequency range as δ decreased (1 mm). Nonetheless, the AR bandwidth was below 3 dB, independent of δ, as shown in Figure 7(b).    (Table 1), the simulated impedance bandwidth (|S 11 | < −10 dB) covered the universal UHF-RFID frequency band, as shown in Figure 8(a). e simulated impedance bandwidth and AR bandwidth outside of the optimal W tf (34, 54, 94, and 114 mm) dramatically deteriorated, as shown in Figures 8(a) and 8(b).    (Table 1), the simulated impedance bandwidth (|S 11 | < −10 dB) covered the universal UHF-RFID frequency band, as shown in Figure 9(a). e simulated impedance bandwidth outside of the optimal W bf (34, 54, 94, and 114 mm) shifted to the lower frequency range, while the AR bandwidth deteriorated and became greater than 3 dB, thereby failing to achieve circular polarization, as shown in Figures 9(a) and 9(b).

e Length of Upper Rectangular Patch (L ty ).
e length of upper rectangular patch (L ty ) was optimized to improve the AR bandwidth to cover the universal UHF-RFID frequency band. Figures 10(a) (Table 1), the simulated impedance bandwidth (|S 11 | < −10 dB) covered the universal UHF-RFID frequency band, as shown in Figure 10    International Journal of Antennas and Propagation universal UHF-RFID frequency band, and the AR bandwidth deteriorated and became greater than 3 dB, thus failing to achieve circular polarization, as shown in Figures 10(a) and 10(b). Figures 11(a) and 11(b) show the simulated |S 11 | and AR of the proposed antenna under variable lengths of lower rectangular plate (L by ): 57.5, 94.5, 137.5, 177.5, and 217.5 mm. Given the optimal length of lower rectangular patch (L by ) of 137.5 mm (Table 1), the simulated impedance bandwidth (|S 11 | < −10 dB) covered the universal UHF-RFID frequency band, as shown in Figure 11(a). e simulated impedance bandwidth outside of the optimal L by (57.5, 94.5, 177.5, and 217.5 mm) shifted to the lower frequency range and deteriorated, while the AR bandwidth deteriorated and became greater than 3 dB, thus failing to achieve circular polarization, as shown in Figures 11(a) and 11(b). Figure 12 illustrates the simulated vector current distribution on the upper-layer diagonally aligned rectangular plates at the center frequency of 900 MHz relative to time (t). e current from the feeding point at the center of the upper-layer plates was distributed on the upper-layer conductor surface with the same magnitude and 90°phase difference, independent of time (t). e direction of surface current on the upper-layer diagonally aligned rectangular plates was counterclockwise relative to time (t). Meanwhile, the surface current on the lower-layer diagonally adjoined rectangular plates was clockwise relative to time (t). e directions of surface current on the upper-and lower-   International Journal of Antennas and Propagation layer rectangular plates at 840 and 960 MHz resembled those at the center frequency. e concurrent use of upper-and lower-layer diagonally aligned rectangular plates thus resulted in circular polarization. Figure 13 shows the simulated current distribution on one end of the wall patch at the center frequency of 900 MHz relative to time (t).

Vector Current Distribution on the Upper-Layer Diagonally Aligned Rectangular Plates.
e direction of surface current was counterclockwise, while that of the other end of the wall patch was clockwise. e opposite directions of surface current resulted in circular polarization. e directions of surface current on the both ends of the wall patch at 840 and 960 MHz resembled those at the center frequency.

Simulation and Measurement Results
A prototype antenna was fabricated based on the optimal dimensions (Table 1), and far-field experiments were carried out in an anechoic chamber using HP8720C network analyzer. Figure 14 depicts the prototype broadband CP bidirectional antenna. Figure 15(a) illustrates the simulated and measured |S 11 | of the broadband CP bidirectional antenna. e simulated and measured |S 11 | (≤−10 dB) covered the frequency range of 772.19 -1014.6 MHz (27.13% bandwidth) and 759 -1011 MHz (28.47% bandwidth), respectively, both of which encompassed the universal UHF-RFID frequency band of 840 -960 MHz. e simulated and measured results were in good agreement. In Figure 15(b), the simulated and measured 3-dB AR bandwidth of the proposed antenna were 675 -1000 MHz (38.80% AR bandwidth) and 648 -1110 MHz (52.55% AR bandwidth), respectively. Figures 16(a)-16(c) show the simulated and measured x-z and y-z plane radiation patterns at 840, 900 (center frequency), and 960 MHz, respectively. e total radiation patterns were bidirectional in both planes with circular polarization. e radiation patterns were near-symmetrical in front-back boresight for the universal UHF-RFID frequency band. e CP pattern is LHCP in direction of 180°along the operating frequency band. e simulated LHCP half-power beamwidth (HPBW) in x-z and y-z planes at 840, 900, and   960 MHz was 90°and 56°; 86°and 56°; and 82°and 56°. e corresponding measured LHCP HPBW were 98°and 52°; 97°a nd 52°; and 88°and 56°. e RHCP pattern occurred in direction of 0°. e RHCP HPBW in x-z and y-z planes at 840, 900, and 960 MHz were 92°and 56°; 84°and 54°; and 76°a nd 54°. e corresponding measured RHCP HPBW were 79°a nd 62°; 97°and 63°; and 88°and 65°. e simulated xz and yz plane 3-dB AR beamwidth at 840, 900, and 960 MHz was 76°and 60°; 80°and 62°; and 104°and 80°. e corresponding measured 3-dB AR beamwidth was 123°a nd 118°; 96°and 126°; and 117°and 97°, as shown in Figures 17(a)-17(c), respectively. Figure 18 compares the simulated and measured realized gains of the proposed broadband CP bidirectional antenna. e   Table 2 compares the characteristics of the proposed antenna from simulated and measured results (|S 11 |, AR, gain, HPBW, bandwidth, and pattern) along operating frequency band. e simulated and measured results are in good agreement.

Comparison between CP Bidirectional Antennas
e overall dimension of the proposed single-fed broadband CP bidirectional antenna is 146 mm × 275 mm × 10 mm excluding the coaxial cable and N-type connector. e coaxial cable and the connector were required to feed radio frequency (RF) signal to the upper-layer diagonally aligned rectangular plates, and the length of the coaxial cable was λ/4 at the center frequency. Table 3 compares the proposed broadband CP bidirectional antenna with existing CP bidirectional antennas. In [6,12], the higher-frequency antennas achieved improved impedance and AR bandwidths but lower antenna gain. In [7], the antenna achieved higher gain but narrower impedance and AR bandwidths. In [8,13], the antennas were smaller than the proposed antenna but achieved lower antenna gain. In [9], the antenna had narrower AR bandwidth and lower antenna gain. In [11,14], the antennas achieved higher gain but suffered from bulkiness. For antennas operable in the universal UHF-RFID frequency band (840-960 MHz), the proposed single-fed broadband CP bidirectional antenna efficiently achieves high bidirectional antenna gain with compact size.

Conclusions
is experiment research introduced a technique of diagonal alignment of two rectangular copper plates to generate circular polarization and improve 3-dB AR bandwidth, where the principle of circular polarization is proved by surface current analysis. Moreover, double layers of diagonally aligned rectangular plates were adopted to improve impedance bandwidth (S 11 | < −10 dB) and achieve bidirectional radiation pattern. According to the characteristics that were analyzed and described in detail of the parametric study section, the prototype antenna was fabricated based on the optimal dimensions, and far-field experiments were carried out in an anechoic chamber using HP8720C network analyzer. e impedance bandwidth, 3-dB AR bandwidth, LHCP/ RHCP HPBW, 3-dB AR beamwidth, and gain of the prototype UHF-RFID antenna were 759 -1011 MHz (28.74%), 648 -1110 MHz (52.55%), 52°-98°/62°-97°, 96°-126°, and 4.28 -5.72 dBic, respectively. e proposed single-fed broadband CP bidirectional antenna is thus universally applicable to UHF-RFID readers of varying geographical areas.

Data Availability
e data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest
e authors declare that there are no conflicts of interest regarding the publication of this paper.