A Novel 3D PILA-Type UHF RFID Tag Antenna Mountable on Metallic Objects for IoT Indoor Localization

In this paper, a novel 3D planar inverted-L antenna (PILA) Ultrahigh Frequency (UHF) Radio Frequency Identification (RFID) tag mountable on metallic surfaces is proposed for the Internet of Things (IoT) indoor localization applications. The proposed tag antenna (45 mm × 82 mm × 4 mm or 0.137λ × 0.25λ × 0.012λ) is designed for mounting on metallic objects. The 3D PILA antenna is fabricated using a copper sheet of thickness 1 mm and air as the dielectric substrate in order to minimize costs for materials and realization. In the design, T-slot has been inserted in the radiating element for tuning of the tag’s resonance for achieving good matching with the chip. Also, a simple equivalent circuit model has been obtained to analyze the impedance of the 3D PILA. Based on the optimized design, the fabricated prototype has been measured in the anechoic chamber. The resonant frequency of the proposed tag is stable, and it is not affected much by the metallic object. The measurement results of the antenna prototype demonstrated a reasonable agreement with the simulation results, and a read range of 3.6 m was measured inside an anechoic chamber. Most importantly, in the building hallway, the proposed tag is able to achieve a maximum read distance of 18 m with a transmitted power of 31.5 dBm at 867 MHz when placed on metal. With the 3D PILA antenna structure, the proposed antimetal tag is a suitable solution that can be integrated into an indoor localization scenario.


Introduction
Indoor localization systems have attracted research interest during the last decade due to their various application [1]. However, indoor positioning techniques are among the most challenging issues related to the Internet of ings (IoT). IoT system is the result of the integration of several technologies, including sensing and methods for identification. Radio Frequency Identification (RFID) and wireless sensor networks (WSNs) are the most important blocks in the framework of the IoT [2]. In IoT, RFID technology is used in smart cities applications such as cameras, human tracking, smart parking, GPS (global positioning system), and intelligent sensors. e RFID system is an identification technology, automatically and in real time, capturing and transmitting data to identify objects through reflected radio waves [3,4]. However, since UHF RFID tags are directly mounted on or near metallic objects, the presence of metal objects in contact or close proximity to the tag can essentially degrade the impedance matching, radiation pattern, radiation efficiency, gain, bandwidth, and read range of the tag. In the past few years, several methods have been proposed by different researchers to tackle and overcome this problem by enhancing the tag's performance. In [5,6], inserting a ceramic material between the tag and metallic object has been proposed. However, using ceramics can increase the costs significantly. e artificial magnetic conductor (AMC) ground and an electromagnetic bandgap (EBG) structure are also used for isolating the RFID tags from metal surfaces [7,8]. However, applications of the AMC and EBG structures can help improve the gain and the read range, but they introduce higher fabrication costs and increase the tag size. e other approach is based on the design of planar antennas with an open stub feed for designing on-metal tags [9]. However, the patch resonator with open stub is large in size for antenna design. To achieve miniaturization, metallic vias are used for shorting one edge of the radiating patch to the ground plane to form new antenna structures such as meandered patch antenna with short stub mountable on metallic objects [10]. In recent years, new techniques have been explored for reducing the planar patch size. To achieve compact size, a flexible patch antenna in [11] was designed using the serration technique and was able to achieve read distances of 7 m when attached to a 20 cm × 20 cm metal plate. To further reduce the patch antenna size, the folding method has recently been proposed for designing various flexible RFID tags for on-metal applications [12][13][14][15]. As the size reduces, there is an increase in ohmic losses, consequently affecting the antenna performances such as gain, efficiency, and directive radiation. e planar inverted-F antenna (PIFA) using the metallic surface as ground plane is a popular structure for designing miniature metal tags. By applying the PIFA design concept, the antenna size can be easily reduced by introducing vias [16,17], but the performance of a PIFA is significantly more sensitive to the position of the vias [18]. In this paper, a new type of L-shaped antenna is used for designing a novel 3D tag antenna that can be mounted on a metal object for IoT RFID applications. e proposed 3D PILA tag without substrate does not require the use of vias to achieve the antenna design, which reduces the cost of the manufacturing process. is paper is organized in the following way: in Section 2, the antenna configuration is explained and an equivalent circuit is developed to describe the input impedance of the proposed 3D PILA. In Section 3, experimental results are described. Finally, conclusions are provided in Section 5.

Antenna Structure Design.
e configuration of the PILA used for this study is a well-known antenna because it has many outstanding advantages like simple structure, small size, lightweight, easy method fabrication, and low manufacturing cost. e configuration of the proposed 3D tag antenna is shown in Figure 1. e prototype UHF RFID tag antenna has been realized in an overall volume of 45 × 82 × 4 mm 3 . It consists of a rectangular patch with a T-slot cut at the center. e antenna consists of two-layer thin copper with a thickness of t s � 1 mm, separated by air. To minimize the antenna size, a shorting plate is used to connect the top radiating patch to the ground plane. e shorting stub is in direct contact with one side of the patch.
To design a UHF RFID tag, the impedance (Z a ) of a tag antenna should be conjugate to the chip impedance (Z c ) of the tag. Tag antenna is first simulated and optimized with the electromagnetic (EM) simulation software CST Microwave Studio by monitoring the complex impedance matching, return loss, and tag reading range to better understand the behavior of the tag. If the tag satisfies all of these requirements, the antenna design is done. Otherwise, the design is further modified to achieve a high level of optimization until requirements are met. e microchip selected for integrating with the proposed PILA tag antenna is an Alien Higgs-3 RFID IC (from Alien Corporation's). e equivalent circuit is a parallel RC-circuit with resistance Rp � 1500 Ω and capacitance Cp � 0.85 pF [19]. e antenna is matched to the RFID chip by means of a very simple matching based on a T-slot. e T-slot has been inserted in the radiating element for tuning of the tag's resonance. e slot antenna transforms the capacitive impedance of the PILA antenna to inductive. is slot achieves complex conjugate impedance matching between the antenna and capacitive tag IC for achieving good adaptation matching with the chip to obtain a good gain in the frequency in resonance. Table 1 shows the optimized design parameters of the proposed 3D PILA antenna structure.
However, the matching polarization of the PILA tag and reader antenna affects the communication link. So, both antennas are linearly polarized, and the read performance is dominated by the alignment between the electric fields of the reader and tag antennas. More specifically, our PILA tag should be kept in the right position to be able to be read and written.

Equivalent Circuit Model.
e equivalent circuit model of the proposed 3D PILA tag is derived for analyzing its input impedance. e equivalent circuit model is shown in Figure 2. To build the model, the top radiating patch is represented by the resistance (R a ), inductance (L a ), and capacitance (C a ) in a parallel RLC circuit. As can be seen in Figure 1, the radiator and the ground plane function like a parallel plate capacitor and its capacitance C a [20] can be estimated using the following equation: where A e � WL 3 − W 1 (L 2 + L 1 ) is the effective surface area (by removing the T-slot) of the radiator and ε 0 � 8.854 × 10 − 12 F/m is the permittivity of free space. e resistance R a , which is needed to account for radiation and loss resistances, and inductance L a of the L-shaped patch can be determined using the CurveFit macro model [21]. With reference to Figure 2, the stub is represented by the inductance (L s ) and resistance (R s ) placed in series. By knowing the stub dimension, its inductance L a (in nH) [22,23] can be calculated using the following equation: where t s � 1 mm denotes the copper thickness of the stub. Also from the stub dimension, the resistance of the stub can be estimated using the AC resistance: where x � 2(1 + t s /h)(δ/t s ), δ � 2.2 × 10 −6 m is the skin depth of copper, ρ � 1.72 × 10 −8 Ωm is the copper resistivity, and K c � 2.084 is the current crowding factor [12]. e inductance of the shorting wall is estimated using L sst � 2L s . International Journal of Antennas and Propagation 3 e capacitance (C g ) of the T-slot gap [23] that accommodates the chip can be calculated using the following equation: Furthermore, the antenna input impedance can be calculated by combining circuit components in the equivalent circuit, and the results are plotted in Figure 3. e input impedance of the proposed 3D PILA generated from the CST simulation is also added to the equivalent circuit model curve for ease of comparison. A slight discrepancy is found between the modeled input impedance and the simulated antenna input impedance. In reality, the approximation equations used for PILA antenna to calculate the element values are derived by assuming uniform current distribution on the antenna. e tag antenna should be matched to the measured impedance of the RFID ASIC chip Z measurement c � (26 − j163)Ω [24] in order to achieve conjugate matching. Indeed, at 900 MHz, the modeled input impedance of the antenna is Z modeled ant � (22.5 − j180)Ω, and the simulated impedance of the antenna is Z simulated ant � (12.5 − j164)Ω. However, despite minor divergences, a reasonable agreement is observed and it shows that the proposed equivalent circuit model is valid to be used for the impedance analysis.  Parameter International Journal of Antennas and Propagation

Results and Discussion
e proposed 3D PILA tag is placed on the center of a piece of 300 × 300 mm 2 metallic plate in simulations and measurements. Figure 4 shows the simulated and measured reflection coefficients of the proposed tag antenna. e simulated return loss of 20 dB is obtained at the frequency of 907 MHz without the metallic plate. Good impedance matching was achieved at 900 MHz for a 17.7 dB return loss with the metallic plate. It could be observed that the resonant frequency of the tag is stable and it is not affected much by the metallic object. By simply changing the length of the matching T-slot onto a radiating element, the return loss of the antenna can be tuned to match the impedance value of the chip. e height of the tag (h) is an important parameter in terms of resonant frequency, bandwidth, and antenna gain. By increasing the height of the top plate, the bandwidth is increased but the reflection coefficient is decreased. e quality factor Q is determined by the inverse of antenna bandwidth. Hence, the best way to increase antenna bandwidth (low Q) is to maximize the volume of the proposed 3D PILA tag. So, increasing the antenna area (antenna height) decreases the reflection coefficient. In addition, if we decrease the height of the top plate, the gain of the antenna decreases, and the intensity of the radiation pattern behind the ground plane increases. erefore, the optimum height of the PILA antenna was 4 mm. Figure 5 depicts the simulated 3D radiation patterns at the resonant frequency of the prototype PILA antenna in free space and on a 30 × 30 cm 2 metal plate. Simulation results show that a tag antenna without a metallic plate has gain 1.98 dB at 905 MHz and an antenna with a metallic plate has gain 4.08 dB at 900 MHz. erefore, the addition of the metallic plane is found to be useful to improve the antenna gain by strengthening the electromagnetic fields in the boresight direction (+z direction).    Figure 6. It can be observed that when the proposed tag antenna is placed on a 30 × 30 cm 2 metal plate, the direction of maximum radiation is changed. As could be expected, the back radiation of the antenna with metal is reduced with respect to the PILA antenna in free space due to the in-phase reflection properties of the metallic plate.
We characterized the performances of the tag in free space and on the metal plate. A procedure for the measurement of our PILA antenna is presented in [25]. Figure 7 shows the experimental setup in the anechoic chamber. In the measurement process, the tag and the horn antenna (reader) are placed face-to-face on a straight horizontal plane.
e distance between the reader antenna and the PILA tag antenna is fixed to d � 55 cm, while the transmitted power sent by the reader is varied gradually until the tag is activated. e tag is placed at the middle of a metal plate in the xy-plane with angle φ � 90°. Figure 8 shows the minimum power versus frequency in free space and on a 30 cm × 30 cm metal plate. In free space, the minimum power needed for generating a correct response to a command is approximately 4.5 dBm at 901 MHz. e measured tag sensitivity increases to 9.9 dBm at 897 MHz when the tag is placed on the metal plate, which is less sensitive than that for free space. In the xy-plane (φ � 90°), our tag in free space has a read distance of 3.6 m with the maximum reader output power approximately of 17.1 dBm. Figure 9 shows the test setup for the measurement of the minimum tag turn on power in an anechoic chamber, where the sensitivity is measured by varying vertically the proposed tag along the y-axis. By observing the measured tag sensitivity, it can be seen from Figure 10 that the tag has the best sensitivity of 3.9 dBm at 901 MHz when the distance along the vertical y-axis is equal to 10 cm. erefore, in order to harness the much electromagnetic energy possible from the antenna, the 3D PILA tag must be positioned correctly facing the reader antenna.
To further verify the UHF tag antenna performance in free space and on metal, maximum reading range measurements for the tag were performed in a corridor environment. A procedure for the measurement of the tag is presented in [26,27]. e ingMagic M6e embedded RFID reader module was used with a horn antenna to detect the tag which replies with its identification code. e power delivered varies between 5 and 31.5 dBm (1.4 W) through a computer interface within a frequency range going from 865.6 to 867.6 MHz and the cable losses are about 0.8 dBm.
In order for a passive UHF RFID tag to be successfully read, the Received Signal Strength Indicator (RSSI) is used in order to provide a more accurate tag location. e read range and RSSI are measured in the European (EU) UHF RFID band in the direction of φ � 90°in the xy-plane. Figure 11 shows the measured read range and RSSI when the tag antenna is placed in free space and on the metal plate in the direction of φ � 90°. For free space measurement, the maximum read distance of the tag is measured to be 6.3 m with 31.5 dBm transmit power and the measured receiver sensitivity is −69 dBm at the frequency of 867 MHz. As depicted in Figure 11, the tag is able to achieve a maximum reading distance of ∼9.6 m when it is tested on a 30 × 30 cm 2 metallic plate, at the frequency of 867.5 MHz. However, when measuring the read performance in the xy plane, the PILA tag antenna is rotated 45 o clockwise about its own zaxis. e read range (φ � 45°) increases to 11 m with RSSI � −72 dBm at 866.9 MHz when the tag antenna is placed on a copper plate. Figure 12 shows the measured read distances in the boresight direction (θ � 0°, φ � 0°) in the xyplane.
e measured best read range of the tag is 18 m (RSSI � −74 dBm) with 31.5 dBm transmit power at 867 MHz when it is tested on a metallic plate.
Note that the use of ingmagic M6 RFID reader in the 902-928 MHz band (US; FCC) can increase the reading distance of the proposed tag since our tag antenna operates around 900 MHz.
It is clear that our tag works very well in a real environment. However, the reason for this read range increase in corridor environment is attributed to constructive addition of a direct beam and reflected beams, perhaps from the wall and floor, each of which contains a weak indirect signal of the strong direct signal. Consequently, the metal plate is used as a backing reflector to reduce the rear radiation, improve the gain and directivity of the antenna in the front direction, and increase outdoor read ranges. e performance of the proposed 3D PILA is compared with other published metal-mountable UHF RFID tags in Table 2. With reference to the table, the tag in [28] is able to achieve a maximum read distance of 14.6 m in the outdoor environment but the tag size is larger than ours and it requires the use of two FR4 substrates and to be separated by an air layer to increase the bandwidth of the antenna. Similarly, with a large profile, the tag in [26] attains longer read distances than our proposed tag, which is able to reach 46.5 m in the corridor. Moreover, the tag uses a spiral-   shaped artificial magnetic conductor (AMC) to increase the tag gain, which makes the fabrication processes very troublesome. Further, in order to increase the antenna gain and read range, the UHF RFID tag can be associated with a reflector plane as in [29] but this makes the antenna heavy and extremely large. In [30], with a reader transmitting power of 1.2 W, the tag is only able to achieve the maximum read distance of 5 m. In addition, the idea of designing the tag antenna using Hilbert curve structure for tagging large metallic objects is impractical to implement. For the case in [31], the tag has almost the same size as ours with a size 88 mm × 32 mm × 3.2 mm. e tag antenna mounted on the metallic surface is able to achieve a maximum read distance of 7.2 m, but the read performance of the tag in free space has not been studied. Referring to Table 2 again, the tag in [14] is compact and flexible, but employing slots for miniaturization can complicate the design and fabrication process. Also, their achievable read distances inside the anechoic cabinet are shorter than ours. However, the read distances are also estimated in [14], which is able to reach 5.2 m, when the tag   is placed on a metal plate. With a reasonable tag size, our proposed tag is able to achieve reasonable read distances in both free space and metallic objects. On the other hand, our tag is made using a copper sheet of a thickness of 1 mm and it does not need the use of the substrate. In addition, our proposed tag does not require the use of multiple shorting vias, parasitic elements, and meandered slotlines. It is not only much simple to design but it can also be easily manufactured, which is preferable for tracking metal objects of various IoT environments in indoor positioning systems. erefore, our 3D PILA tag can easily be used in heavy industry and outdoor applications.

Manufacturing Process
e manufacturing process begins with sourcing all the parts, namely, copper sheet and plastic support. e second step in the process is to source the consumables such as adhesive and solder. Indeed, the antenna elements were made with a 1 mm copper sheet and the substrate of the whole proposed antenna is air. Plastic support is used in between the radiating element (patch antenna) and the ground plane. e next step is cutting the copper sheet by laser to obtain the adjacent antenna parts (Figure 1(c)). After cutting the antenna to the desired length, it will be folded to have the final 3D structure of our tag. Furthermore, soldering the IC to the structure and inserting the plastic backing constitute the final manufacturing step. However, the total cost for manufacturing per tag is no more than $1 (USD).

Conclusion and Future Work
A novel design of a 3D PILA tag antenna for the UHF RFID system has been proposed for mounting on metallic objects. It is shown that by using two copper plates and separated by an air layer, the structure is insensitive to the presence of metallic objects, its gain can be increased, and the bandwidth covers all passive UHF RFID band. e T-slot is also applied to the design for adjusting the resistance and reactance of the antenna impedance. A simple and cost-effective method for fabrication is possible with this design while eliminating the need for vias and complicated multilayer structures. An equivalent circuit has been developed for analyzing the impedance characteristics of the tag antenna. Good agreement is found between the simulation and modeled data. Simulations and measurements have been conducted to demonstrate the tag performances to metallic and nonmetallic objects. e maximum read ranges of our proposed tag are 6.3 m in free space and 18 m when mounted on metal surfaces under real operating conditions (building hallway). When compared with some metal tags, our proposed tag is simple in structure and it is able to achieve reasonable read distances in free space and on metallic objects. e proposed antenna is a good candidate for RFID and IoTapplications in real environments.
In future, we can design the Tag using metamaterial concepts [32][33][34] by placing a metamaterial slab above a PILA antenna (as a superstrate) to focus the signal in one direction to provide a long detection range.

Data Availability
e data used to support the findings of the study are included within the article.