WO2008020658A1

WO2008020658A1 - A printed dipole antenna for rfid tag and the design method therefor

Info

Publication number: WO2008020658A1
Application number: PCT/KR2006/003228
Authority: WO
Inventors: Jun Hwi Choi; Jun Kyung Cho; Gab Young Kim
Original assignee: Hutech21. Co., Ltd.
Priority date: 2006-08-16
Filing date: 2006-08-17
Publication date: 2008-02-21

Abstract

The present invention relates to a printed dipole antenna provided in an RFID tag and a method of designing the same. More particularly, it relates to a printed dipole antenna, wherein a microstrip balun for connecting a printed dipole and a ground plate is disposed on one surface of a substrate and is connected to a feeding line disposed on the opposite surface of the substrate through a via hole passing through the substrate, so that the feeding line and the printed dipole can be effectively matched to each other and distortion of a radiation pattern due to a leakage current can be prevented. The printed dipole antenna for an RFID tag comprises a substrate, a printed dipole including a pair of rectangular arms disposed on a top surface of the substrate, a ground plate disposed on the top surface of the substrate to be spaced apart from the printed dipole, a microstrip balun for connecting each of the arms of the printed dipole to the ground plate, and a feeding line disposed on a bottom surface of the substrate and connected to the microstip balun through a via hole passing through the substrate.

Description

A PRINTED DIPOLE ANTENNA FOR RFID TAG AND THE DESIGN METHOD THEREFOR

Technical Field

[1] The present invention relates to a printed dipole antenna provided in an RFID tag and a method of designing the same. More particularly, the present invention relates to a printed dipole antenna, wherein a microstrip balun for connecting a printed dipole and a ground plate is disposed on one surface of a substrate and is connected to a feeding line disposed on the opposite surface of the substrate through a via hole passing through the substrate, so that the feeding line and the printed dipole can be effectively matched to each other and distortion of a radiation pattern due to a leakage current can be prevented, and a method of designing the printed dipole antenna.

[2]

Background Art

[3] Radio Frequency Identification (RFID) is widely used in a variety of service ind ustries, logistics industries, manufacturing industries and the like. The RFID is operated in such a manner that information stored in a tag is transmitted in a non- contact method from a transponder having an antenna and a microstrip transmitter, i.e. the tag, to a reader corresponding to an information request system.

[4] In a passive tag system, an RF carrier signal transmitted from a reader for the non- contact transmission should provide sufficient power such that a tag can be activated to process data and to retransmit the signal up to a required recognition distance (usually 0.3 to 1 m).

[5] However, since the maximum Effective Isotropic Radiated Power (EIRP) of a reader is restricted by radio regulations, an antenna for a tag with high gain is essentially required to increase the recognition distance.

[6] Recently, a printed dipole antenna has been widely used in the field of an antenna for a tag. Since the printed dipole antenna can be manufactured using a printed circuit technique and thus have superior productivity to a conventional wire dipole antenna, there is an advantage in that the printed dipole antenna is more suitable for mass production.

[7] To enhance matching performance and bandwidth, a balun-integrated wideband printed dipole antenna having an open stub with a 1/4 wavelength is used as the conventional printed dipole antenna.

[8] However, there is a problem in that a coupling effect and power loss may be produced in a transmission line having the open stub with a 1/4 wavelength due to radiation from the open stub.

[9] In the meantime, the conventional printed dipole antenna uses a simulator such as

High Frequency Structure Simulator (HFSS) to derive optimal design variables, thereby improving matching performance, reflection loss, bandwidth or the like.

[10] However, since a simulator having a large amount of operation and a complicated operation process is used in a method of designing the conventional antenna, there is a problem in that excessive time is taken to design the antenna and it is not easy to redesign the antenna for improving its performance.

[H]

Disclosure of Invention Technical Problem

[12] Accordingly, the present invention is conceived to solve the aforementioned problems of the conventional printed dipole antenna. An object of the present invention is to provide a printed dipole antenna, wherein a microstrip balun for connecting a printed dipole and a ground plate is disposed on one surface of a substrate and is connected to a feeding line disposed on the opposite surface of the substrate through a via hole passing through the substrate, so that the feeding line and the printed dipole can be effectively matched to each other and distortion of a radiation pattern due to a leakage current can be prevented, whereby an enhanced bandwidth, a radiation pattern in a single direction and high antenna gain can be obtained.

[13] Another object of the present invention is to provide a method of designing an antenna, wherein an artificial neural network (ANN) model of an antenna to be designed is established and design variables capable of obtaining optimized performance are calculated at a rapid speed using the established ANN model, so that accuracy can be ensured and a period of time needed to design an antenna can be remarkably reduced.

[14]

Technical Solution

[15] According to an aspect of the present invention for achieving the objects, there is provided a printed dipole antenna for an RFID tag, which comprises a substrate, a printed dipole including a pair of rectangular arms disposed on a top surface of the substrate, a ground plate disposed on the top surface of the substrate to be spaced apart from the printed dipole, a microstrip balun for connecting each of the arms of the printed dipole to the ground plate, and a feeding line disposed on a bottom surface of the substrate and connected to the microstip balun through a via hole passing through the substrate.

[16] According to another aspect of the present invention, there is provided a method of designing a printed dipole antenna for an RFID tag including a microstrip balun for connecting a printed dipole and a feeding line through a via hole, which comprises the steps of selecting design variables and performance variables to be optimized, establishing an artificial neural network (ANN) model of the antenna, and calculating design variables capable of obtaining optimal antenna characteristics using the established ANN model. [17]

Advantageous Effects

[18] A printed dipole antenna for an RFID tag according to the present invention is provided with a microstrip balun connected to a feeding line through a via hole, so that matching performance can be further enhanced and distortion of a radiation pattern due to a leakage current flowing through an outer conductor of a coaxial line can be prevented.

[19] Further, a method of designing a printed dipole antenna for an RFID tag employs an artificial neural network (ANN) model, so that accuracy close to EM simulation can be ensured and a calculation time can also be remarkably reduced to thereby reduce costs and time required in the antenna design.

[20]

Brief Description of the Drawings

[21] Fig. 1 is a perspective view of a printed dipole antenna according to an embodiment of the present invention.

[22] Fig. 2 is a plan view of the printed dipole antenna shown in Fig. 1.

[23] Fig. 3 is a bottom view of the printed dipole antenna shown in Fig. 1.

[24] Fig. 4 is a view showing an equivalent circuit model for the printed dipole antenna shown in Fig. 1.

[25] Fig. 5 is a flowchart illustrating a method of designing a printed dipole antenna according to an embodiment of the present invention.

[26] Fig. 6 is a view showing a structure of an artificial neural network (ANN) model applied to a step of establishing the ANN model shown in Fig. 5.

[27] Fig. 7 is a flowchart illustrating in detail the step of establishing the ANN model shown in Fig. 5.

[28] Fig. 8 is a view illustrating results in which an output of the ANN model established in accordance with the step of establishing the ANN model shown in Fig. 7 and an output obtained from EM simulation are compared with each other.

[29] [Explanation of Reference Numerals for Major Portions Shown in Drawings]

[30] 10: Printed dipole 20: Microstrip balun

[31] 25 : Via hole 30: Ground plate [32] 40: Substrate 50: Feeding line

[33]

Best Mode for Carrying Out the Invention

[34] Hereinafter, a preferred embodiment of the present invention will be described in detail with reference to the accompanying drawings.

[35] Fig. 1 is a perspective view of a printed dipole antenna for an RFID tag according to an embodiment of the present invention; and Figs. 2 and 3 are plan and bottom views of the printed dipole antenna for an RFID tag shown in Fig. 1, respectively.

[36] As shown in Figs. 1 to 3, the printed dipole antenna for an RFID tag according to an embodiment of the present invention comprises a printed dipole 10 including a pair of rectangular arms disposed on a top surface of a substrate 40, a ground plate 30 disposed on the same surface of the substrate 40 as the printed dipole 10 to be spaced apart from the printed dipole 10, a microstrip balun 20 for connecting each of the arms of the printed dipole 10 to the ground plate 30, and a feeding line 50 disposed on a bottom surface of the substrate 40 and connected to the microstip balun 20 through a via hole 25 passing through the substrate 40.

[37] The microstrip balun 20 is connected to the feeding line 50 by means of a coaxial line (not shown) passing through the via hole 25 and serves to allow the coaxial line and the printed dipole 10 to be matched to each other and to prevent distortion of a radiation pattern which may be produced due to a leakage current flowing through an outer conductor of the coaxial line.

[38] Here, the lengths of the microstip balun 20 and the respective arms of the printed dipole 10 is preferably close to a 1/4 wavelength.

[39] Further, the substrate 40 is preferably formed of FR4 that is relatively inexpensive and has a high dielectric constant in a region of 10 GHz or less. Furthermore, FR4 with a thickness of 0.8 to 1 mm, a relative permittivity of 4.5 to 5.0 and a loss tangent (tan δ) of 0.014 to 0.015 is preferably used.

[40] Fig. 4 is a view showing an equivalent circuit model for the printed dipole antenna for an RFID tag shown in Fig. 1.

[41] The following expressions for input impedances can be obtained from Fig. 4.

[42]

[43] [Math Figure 1]

[44]

[45]

[46] [Math Figure 2] [47]

Z_inl+j ^■ Z₁ ^• tanCΘQ M^X Z_{1 +}J - Z^₁ - IaHCe₁)

[48]

[49] [Math Figure 3]

[50]

[51] [52] [53] wherein θjand θ₂denote electrical lengths of an impedance converter, θ denotes an b electrical length of a ground line, and θ denotes an electrical length of a slot line. ab

[54] To satisfy feeding equivalence and impedance matching requirements, the respective variables described in Mathematical Expressions (1) to (3) are set as follows. That is, Z = 50 Ω, Z = Z , θ = θ and θ = 90°. θ is also set to be 2.4°

1 in3 1 1 2 ab b which corresponds to an electrical length of a substrate with a thickness of 0.8 mm. [55] Fig. 5 is a flowchart illustrating a method of designing a printed dipole antenna according to an embodiment of the present invention. [56] Hereinafter, a method of designing the aforementioned printed dipole antenna according to an embodiment of the present invention will be described in detail with reference to Fig. 5. [57] First, design variables and performance variables to be optimized are selected

(SlOO). [58] Preferably, a length Ld and a width Wd of a dipole arm, a length Lb of a microstrip balun and a frequency are selected as the design variables, and a magnitude (dB) and phase of reflection loss corresponding to main features of the antenna are selected as the performance variables. [59] Furthermore, other dimensions which have not been selected as input variables are kept constant as described in Table 1. [60] [61] Table 1

[62] [63] Then, an artificial neural network (ANN) model of an antenna to be designed is established (S200). A detailed process of establishing the ANN model will be described later. Here, the structure of the ANN model applied to this process will be described with reference to Fig. 6.

[64] As shown in Fig. 6, the ANN model comprises an input layer, an output layer and a hidden layer. [65] The hidden layer integrates nonlinear functions such that a complex input/output relation existing between a plurality of inputs and outputs can be modeled. [66] The input layer is connected to the hidden layer by a group of weight sets, and the hidden layer is also connected to the output layer by another group of weight sets.

[67] Such weights are adjusted until a desired response is obtained through the training. [68] After a step of establishing the ANN model has been completed, design variables with which optimal antenna characteristics can be obtained are calculated using the established ANN model (S300). Therefore, the design for the printed dipole antenna can be finished.

[69] Hereinafter, a detailed procedure for the step of establishing the ANN model will be described in detail with reference to Fig. 7. [70] First, output data for a specific value of a design variable within a certain range are obtained using a previously verified tool and a training data set including input and output data is then acquired (S210).

[71] Preferably, the input and output data are normalized to have a value between -1 and 1, so that neuron saturation cannot be produced during the training of the ANN model due to excessive increase in an activation value.

[72] Further, Agilent Momentum EM simulation (hereinafter referred to as "EM simulation") is preferably used as the previously verified tool. [73] That is, EM simulation for a plurality of sample points within the ranges of input parameters shown in Table 2 can be performed in a frequency range of 2 to 3 GHz.

[74] [75] Table 2

[76] [77] Then, the selected design variables and performance variables are assigned to neurons in the input and output layers, respectively, and the number of neurons in the hidden layer is specified to thereby set neurons of the ANN model and initial weights (S220).

[78] Here, the number of neurons in the hidden layer is preferably set to five in consideration of convergence rate and accuracy. [79] Then, an input vector obtained from the acquired training data set is assigned to an input neuron in the input layer, and an output vector in the output layer is calculated using the initial weight set (S230).

[80] Thereafter, an output value of the calculated output vector is compared with the output data of the training date set to thereby calculate a training error (S240). [81] Subsequently, the calculated training error is compared with the predetermined critical value (S250). If the training error is greater than the critical value, each of the weights of the ANN model are reset (S260) and a process is returned to a step S230 of calculating the output vector to repeat the steps S230 to S250. If the training error is equal to or lower than the critical value, the establishment of the ANN model is completed, and the step S300 of calculating the optimal design variables is performed.

[82] If an output value obtained using the ANN model so established is compared with an output value obtained from the EM simulation, it can be seen that both of the magnitude and phase of reflection loss are almost the same as shown in Fig. 8 (a) and (b). Here, the input values are set such that L = 21.4 mm, W = 6 mm and L = 23 mm. d d b

[83] Accordingly, a well established ANN model can be used instead of the EM simulation to optimally design an antenna, and a time taken to calculate the design variable can be remarkably reduced by using the ANN model.

[84] Required time has been compared using a PC (2.8 GHz CPU and 256 MB RAM). As a result, about 10 minutes is required in the EM simulation, whereas about 10 seconds is required in the ANN model driven by Matlab.

[85] Although the present invention has been described in connection with the ac- companying drawings and the preferred embodiments, the present invention is not limited thereto but defined by the appended claims. Accordingly, it will be understood by those skilled in the art that various modifications and changes can be made thereto without departing from the spirit and scope of the invention defined by the appended claims.

[86]

Industrial Applicability

[87] According to the present invention, there is provided a printed dipole antenna wherein a microstrip balun optimized through a design method with ensured accuracy and reduced design time is disposed. Therefore, an RFID tag system with enhanced bandwidth and high antenna gain can be manufactured.

Claims

[1] A printed dipole antenna for an RFID tag, comprising: a substrate; a printed dipole including a pair of rectangular arms disposed on a top surface of the substrate; a ground plate disposed on the top surface of the substrate to be spaced apart from the printed dipole; a microstrip balun for connecting each of the arms of the printed dipole to the ground plate; and a feeding line disposed on a bottom surface of the substrate and connected to the microstip balun through a via hole passing through the substrate. [2] The printed dipole antenna as claimed in claim 1, wherein the microstrip balun and each of the arms of the printed dipole have lengths close to a 1/4 wavelength. [3] The printed dipole antenna as claimed in claim 1, wherein the substrate is formed of FR4 with a thickness of 0.8 to 1 mm, a relative permittivity of 4.5 to 5.0 and a loss tangent (tan δ) of 0.014 to 0.015. [4] A method of designing a printed dipole antenna for an RFID tag including a microstrip balun for connecting a printed dipole and a feeding line through a via hole, comprising the steps of:

(a) selecting design variables and performance variables to be optimized;

(b) establishing an artificial neural network (ANN) model of the antenna; and

(c) calculating design variables capable of obtaining optimal antenna characteristics using the established ANN model.

[5] The method as claimed in claim 4, wherein step (a) comprises the step of selecting a length (L d ) and width (W d ) of a dipole arm, a length (L b ) of a microstrip balun and a frequency as the design variables, and selecting a magnitude (dB) and phase of reflection loss as the performance variables. [6] The method as claimed in claim 4, wherein step (b) comprises the steps of:

(bl) obtaining output data for a specific value of a design variable within a certain range using a previously verified tool, and acquiring a training data set including input and output data;

(b2) assigning the selected design variables and performance variables to neurons in input and output layer, respectively, and setting the number of neurons in a hidden layer and initial weights;

(b3) assigning an input vector obtained from the acquired training data set to an input neuron in the input layer, and calculating an output vector in the output layer using a set of the initial weights;

(b4) comparing an output value of the calculated output vector and output data of the training data set to thereby calculate a training error;

(b5) comparing the calculated training error and a predetermined critical value; and

(b6) resetting each of the weights in the ANN model and returning to the output vector calculation step, if it is determined in the comparison step that the training error is greater than the critical value. [7] The method as claimed in claim 6, wherein step (bl) comprises the step of normalizing the input and output data to have a value between -1 and 1. [8] The method as claimed in claim 6, wherein step (bl) comprises the step of using

Agilent Momentum EM simulation as the previously verified tool. [9] The method as claimed in claim 6, wherein step (b2) comprises the step of setting the number of neurons in the hidden layer to five.