Mutual Coupling Reduction in Two-Dimensional Array of Microstrip Antennas Using Concave Rectangular Patches
JJOURNAL OF TELECOMMUNICATIONS, VOLUME 2, ISSUE 2, MAY 2010 64
Mutual Coupling Reduction in Two-Dimensional Array of Microstrip Antennas Using Concave Rectangular Patches
Shahram Mohanna, Ali Farahbakhsh, and Saeed Tavakoli
Abstract — Using concave rectangular patches, a new solution to reduce mutual coupling and return loss in two-dimensional array of microstrip antennas is proposed. The effect of width and length concavity on mutual coupling and return loss is studied. Also, the patch parameters as well as the amounts of width and length concavity are optimized using an enhanced genetic algorithm. Simulation results show that the resulting array antenna has low amounts of mutual coupling and return loss.
Index Terms — Concave patch, enhanced genetic algorithm, microstrip array antenna, mutual coupling, optimization, return loss. —————————— (cid:1) ——————————
1 I
NTRODUCTION
ICROSTRIP array antennas are used widely be-cause of their simple manufacturing, small size, light weight and low cost [1-6]. They are used in phased array antennas applications such as pattern beam forming, smart antennas, and electronic scanning radars [4, 5]. To calculate the radiation pattern of an array the mutual coupling effect must be considered. Without con-sidering this effect, large errors in beam forming and null string are resulted [7, 8]. To decrease the mutual coupling effect, several methods such as changing feed position and feed structure and replacing ordinary patches with new types, such as fractal patches, have been reported [9-14]. To reduce mutual coupling, the use of concave rectan-gular patches is proposed. In [9], this method was applied to a linear array. In this paper, the patch concavity effect on a two-dimensional array, including four elements, is studied. In addition, the effect of patch concavity on re-turn loss is investigated. Furthermore, the patch length and width as well as the amount of width and length con-cavity are optimized employing an enhanced genetic al-gorithm.
2 S
TRUCTURE OF T HE A RRAY A NTENNA
The array antenna includes four elements, which all of them are implemented on one antenna layer structure. The antenna layer structure includes three conductive and two dielectric layers. The lower conductive layer is assumed to be an infinite ground plane, while the second one is set to be a feed layer and the patches are placed in the top layer. In simulation, the thickness of dielectric layers, D and D , are set to 0.5mm. The dielectric relative permittivity, r ε , and the dielectric relative permeability, r µ , are equal to 3 and 1, respectively. Figure 1 shows the side view of the antenna. Fig. 1. Layer structure of the antenna
Figure 2 illustrates the array structure and its dimen-sions. The patch dimensions can be specified by the fre-quency and the bandwidth of the antenna [6]. In this study, the antenna is designed for X band applications. The patch width and length is set to be mmW = and mmL = , results the resonant frequency to be 8.55GHz. The patches are fed by proximity coupled me-chanism with L shape feed lines, having a line impedance of 50 Ohm. As shown in Figure 2, the distances between center of elements in X and Y directions are 14.08mm and 20.55mm, respectively.
3 P
ATCH C ONCAVITY E FFECT
Figure 3 illustrates the concavity on the above-mentioned array antenna, where h and h refer to the depths of the width concavity and length concavity, respectively. In this section, the effect of concavity in width, length, and both sides on the mutual coupling and reflection coeffi-cient is studied. Employing FEKO software [15], antenna parameters are obtained based on the method of mo-ments [16]. M ———————————————— • S. Mohanna is with the Faculty of Electrical and Computer Engineering, The University of Sistan and Baluchestan, Zahedan, Iran. • A. Farahbakhsh is with the Faculty of Electrical and Computer Engineer-ing, The University of Sistan and Baluchestan, Zahedan, Iran. • S. Tavakoli is with the Faculty of Electrical and Computer Engineering, The University of Sistan and Baluchestan, Zahedan, Iran. © 2010 JOT© 2010 JOT
Figure 3 illustrates the concavity on the above-mentioned array antenna, where h and h refer to the depths of the width concavity and length concavity, respectively. In this section, the effect of concavity in width, length, and both sides on the mutual coupling and reflection coeffi-cient is studied. Employing FEKO software [15], antenna parameters are obtained based on the method of mo-ments [16]. M ———————————————— • S. Mohanna is with the Faculty of Electrical and Computer Engineering, The University of Sistan and Baluchestan, Zahedan, Iran. • A. Farahbakhsh is with the Faculty of Electrical and Computer Engineer-ing, The University of Sistan and Baluchestan, Zahedan, Iran. • S. Tavakoli is with the Faculty of Electrical and Computer Engineering, The University of Sistan and Baluchestan, Zahedan, Iran. © 2010 JOT© 2010 JOT http://sites.google.com/site/journaloftelecommunications/
OURNAL OF TELECOMMUNICATIONS, VOLUME 2, ISSUE 2, MAY 2010 65
The effect of the width concavity of the patches ( h ) on the mutual coupling and reflection coefficient is investi-gated. The mutual coupling and reflection coefficient are computed for mmh = . Fig. 2. Structure of array antenna used in the simulation
Fig. 3. Structure of concave array antenna
The patch with no concavity is shown by mmh = . The scattering parameters, which are related to either the reflection coefficient or mutual coupling, are depicted in Figures 4 and 5. It can be seen that the reflection coeffi-cient is increased and the mutual coupling is decreased by increasing the amount of concavity. Although the width concavity can decrease the effect of mutual coupl-ing, the resonant frequency is shifted up. To compensate this effect and move the resonant frequency back to its initial value, patch dimensions should be increased. The effect of the length concavity of the patches ( h ) on mutual coupling and reflection coefficient is studied. Fig-ures 6 and Figure 7 show the scattering parameters. It is seen that the reflection coefficient is minimized in mmh = . By increasing the amount of the length concav-ity, parameters S and S are decreased, whereas S is increased. Although the width concavity can decrease the effect of mutual coupling, the resonant frequency is shifted down. To move the resonant frequency back to its initial value, patch dimensions should be decreased. This effect is useful in designing small-size antennas. Fig. 4. Reflection coefficient for width concavity
Fig. 5. Coefficients related to mutual coupling for width concavity © 2010 JOT http://sites.google.com/site/journaloftelecommunications/
OURNAL OF TELECOMMUNICATIONS, VOLUME 2, ISSUE 2, MAY 2010 66
Fig. 6. Reflection coefficient for length concavity
Fig. 7. Coefficients related to mutual coupling for length concavity
For simplicity the depths of concavity in both sides, h and h , are considered to be the same. The effect of the two-side concavity on the reflection coefficient and mu-tual coupling are shown in Figures 8 and 9. By increasing the amount of concavity, it can be seen that the reflection coefficient is increased, whereas the mutual coupling is decreased. Like the previous cases, the resonant frequen-cy is changed. From all three cases considered in this section, it can be concluded that the concavity can influence the mutual coupling effect and reflection coefficient. Therefore, by changing the size of width and length as well as the amount of concavity, an array antenna with a desired resonant frequency and low amounts of mutual coupling and reflection coefficient can be designed. In addition, the parameter showing the amount of feed line and patch overlap affects antenna properties. Therefore, this para-meter should also be considered in the design procedure. As this procedure is based on trial and error, it is time-consuming. More importantly, if the concavity in each side is considered to be different, the trial and error pro-cedure will be more complicated. Therefore, it is pro-posed to employ an optimization procedure to determine the optimal values of the patch length, the patch width, the amount of overlap, as well as the amounts of width and length concavity.
4 E
NHANCED G ENETIC A LGORITHM
Based on an analogy to the phenomenon of natural selec-tion in biology, Holland [17] proposed the GA method. First, the optimization problem is given a chromosome structure. Next, an initial population is generated, ran-domly. Then, members of the population with higher fit-ness are selected. The fitness of members is calculated by an evaluation function. A member with a higher fitness has a more chance to be selected; therefore, weaker mem-bers are gradually replaced by stronger members. Se-lected members mate two by two randomly and the next population is generated. This procedure is repeated until the stop condition is reached. To prevent the algorithm to converge to local optimums, the mutation operator, which generates new chromosomes with different charac-teristics, is also applied. The mutation and crossover op-erators can be interpreted as negative and positive feed-backs, respectively. The coefficient of crossover acts as a positive feedback. When crossover coefficient is increased, the algorithm is forced to converge. If it is increased too much, the algo-rithm may converge to a local optimum. The mutation coefficient plays the role of the negative feedback. In-creasing mutation coefficient results in a deep search but at the cost of a low-speed convergence. In contrast to conventional GA, in which coefficients of feedbacks are constant during running the optimization process, the method presented in [18] suggests changing these coefficients, using fuzzy systems. This leads to a good trade-off between the speed of algorithm and the accuracy of the solution. © 2010 JOT© 2010 JOT
Based on an analogy to the phenomenon of natural selec-tion in biology, Holland [17] proposed the GA method. First, the optimization problem is given a chromosome structure. Next, an initial population is generated, ran-domly. Then, members of the population with higher fit-ness are selected. The fitness of members is calculated by an evaluation function. A member with a higher fitness has a more chance to be selected; therefore, weaker mem-bers are gradually replaced by stronger members. Se-lected members mate two by two randomly and the next population is generated. This procedure is repeated until the stop condition is reached. To prevent the algorithm to converge to local optimums, the mutation operator, which generates new chromosomes with different charac-teristics, is also applied. The mutation and crossover op-erators can be interpreted as negative and positive feed-backs, respectively. The coefficient of crossover acts as a positive feedback. When crossover coefficient is increased, the algorithm is forced to converge. If it is increased too much, the algo-rithm may converge to a local optimum. The mutation coefficient plays the role of the negative feedback. In-creasing mutation coefficient results in a deep search but at the cost of a low-speed convergence. In contrast to conventional GA, in which coefficients of feedbacks are constant during running the optimization process, the method presented in [18] suggests changing these coefficients, using fuzzy systems. This leads to a good trade-off between the speed of algorithm and the accuracy of the solution. © 2010 JOT© 2010 JOT http://sites.google.com/site/journaloftelecommunications/
OURNAL OF TELECOMMUNICATIONS, VOLUME 2, ISSUE 2, MAY 2010 67
Fig. 8. Reflection coefficient for two-side concavity Fig. 9. Coefficients related to mutual coupling for two-side concavity
To increase the convergence of the algorithm, the posi-tive feedback is increased at the start of the optimization process. Then, the negative feedback is increased to en-sure the globality of the solution. Computing the conver-gence of the algorithm, coefficients of feedbacks are changed appropriately. When the algorithm is about to converge, therefore, the negative feedback should be in-creased. The positive feedback must be increased, when the algorithm fails to converge.
5 O
PTIMIZATION OF A RRAY A NTENNA
To achieve an array antenna with the least amounts of mutual coupling and reflection coefficient, an enhanced genetic algorithm [18] is employed. The optimization pro-cedure aims to minimize the scattering parameters at a resonant frequency of 8.55 GHz. Optimization variables are dimensions of the patches, LW , , the amount of feed line and patch overlap, ins , and the concavity parame-ters, , hh . Based on the above-mentioned optimization procedure, a MATLAB [19] code is developed, which is linked with FEKO. Running this program, the optimal parameters are given by mmW = , mmL = , mmins = , mmh = , and mmh = . The reflection coefficient and mutual coupling are shown in Figures 10 and 11. Fig. 10. Reflection coefficient for the optimal array antenna
6 C
ONCLUSIONS
This paper proposed a new solution to reduce mutual coupling in two-dimensional array of microstrip antennas by using concave patches instead of common rectangular patches. The effect of width and length concavity on the antenna mutual coupling and return loss was studied. Also, an optimal array antenna was designed by employ-ing an enhanced genetic algorithm. Simulation results demonstrated that the resulting array antenna has low amounts of mutual coupling and return loss. © 2010 JOT© 2010 JOT
This paper proposed a new solution to reduce mutual coupling in two-dimensional array of microstrip antennas by using concave patches instead of common rectangular patches. The effect of width and length concavity on the antenna mutual coupling and return loss was studied. Also, an optimal array antenna was designed by employ-ing an enhanced genetic algorithm. Simulation results demonstrated that the resulting array antenna has low amounts of mutual coupling and return loss. © 2010 JOT© 2010 JOT http://sites.google.com/site/journaloftelecommunications/
OURNAL OF TELECOMMUNICATIONS, VOLUME 2, ISSUE 2, MAY 2010 68
Fig. 11. Coefficients related to mutual coupling for the optimal array antenna
References [1]
David M. Pozar, “Microstrip antenna”, Proceeding of the IEEE, Vol. 80, 1992. [2]
K. R. CARVER AND J. W. MINK, “Microstrip Antenna Tech-nology”, IEEE Transactions on Antennas and Propagation, Vol. 29, NO. 1, 1981. [3]
J. P. Gianvittorio and Y. Rahmat-Samii, “Fractal antennas: A novel antenna miniaturization technique and applications,” IEEE Antennas Propagation. Mag., vol. 44, pp. 20–36, 2002. [4]
R. Osman, “Microstrip array antenna for automobile radar system”, MSc thesis, University Technology Malaysia, 2006. [5]
Robert J. Mailloux, “Phased Array Antenna Handbook”, AR-TECH HOUSE, 2005 [6]
Constantine A. Balanis, “Modern Antenna Handbook”, Wiley, 2008 [7]
K.R. Dandekar, H. Ling and G. Xu, "Effect of mutual coupling on direction finding in smart antenna applications", Electronics Letters, Vol. 36, 2000. [8]
R. Bhagavatula, R. W. Heath Jr., A. Forenza, N. J. Kirsch and K. R. Dandekar, “Impact of mutual coupling on adaptive switch-ing between MIMO transmission strategies and antenna confi-gurations”, Springer Science and Business Media, 2008. [9]
A.Farahbaksh, S. Mohanna and S. Tavakoli, “Reduction of mu-tual coupling in microstrip array antennas using concave rec-tangular patches”, 2009 international symposium on Antennas and propagation, Bangkok, Thailand, October 2009 [10]
A. Farahbakhsh, S. Mohanna, S. Tavakoli and M. Oukati Sa-degh, “New Patch Configurations to Reduce the Mutual Coupl-ing in Microstrip Array Antennas”, Antennas and Propagation Conference, Loughborough, England, November 2009. [11]
M. M. Nikolic´, A. R. Djordjevic´and A. Nehorai, “Microstrip antennas with suppressed radiation in horizontal directions and reduced coupling” IEEE Transactions on Antennas and Propagation, Vol. 53, No. 11, 2005. [12]
K. Parthasarathy, “Mutual coupling in patch antenna arrays”, MSc thesis, University of Cincinnati, 2006. [13]
N. Yousefzadeh and C. Ghobadi, “Decreasing of Mutual Coupl-ing in Array Antennas by Using Fractal Elements”, Progress In Electromagnetics Research Symposium 2007, Beijing, China, March [14]
S. Jarchi, J. Rashed-Mohassel and M. H. Neshati, “Mutual Coupling of Rectangular DRA in a Four Element Circular Ar-ray”, Progress In Electromagnetics Research Symposium 2007, Beijing, China, March [15]
FEKO 5.4, Copyright 2005-2008, EM Software & Systems-S.A. (Pty) Ltd. [16]
Ney, M. M., “Method of moments as applied to electromagnetic problems,” IEEE Transaction on Microwave Theory Tech., Vol. 33, No. 10, 972–980. 1985. [17]
Holland JH. Adaptation in natural and artificial systems. 1975, Ann Arbor: University of Michigan Press. [18]
A. Farahbakhsh, S. Tavakoli and A. Seifolhosseini, “Enhance-ment of genetic algorithm and ant colony optimization tech-niques using fuzzy systems”, IEEE International Advance Computing Conference, March 2009. [19]
MATLAB 7.1, Copyright 1984-2005, The MathWorks Inc.
Shahram Mohanna received his BSc and MSc degrees in electrical engineering from the University of Sistan and Baluchestan, Iran and the University of Shiraz, Iran in 1990 and 1994, respectively. He then joined the University of Sistan and Baluchestan, Iran. In 2005, he obtained his PhD degree in electrical engineering from the University of Manchester, England. As an assistant professor at the University of Sistan and Baluchestan, his areas of research include design of microwave circuits, antenna design and applied electromagnetic. Dr. Mohanna has served as a reviewer for several journals and a num-ber of conferences . Ali Farahbakhsh obtained his BSc and MSc degrees in electrical engineering from Shahid Bahonar University of Kerman, Iran and the University of Sistan and Baluchestan, Iran in 2007 and 2010, respec-tively. His main area of research is design of microstrip array anten-nas.
Saeed Tavakoli received his BSc and MSc degrees in electrical engineering from Ferdowsi University of Mashhad, Iran in 1991 and 1995, respectively. In 1995, he joined the University of Sistan and Baluchestan, Iran. He earned his PhD degree in electrical engineer-ing from the University of Sheffield, England in 2005. As an assistant professor at the University of Sistan and Baluchestan, his research © 2010 JOT http://sites.google.com/site/journaloftelecommunications/
OURNAL OF TELECOMMUNICATIONS, VOLUME 2, ISSUE 2, MAY 2010 69 interests are space mapping optimization, multi-objective optimiza-tion, control of time delay systems, PID control design, robust con-trol, and jet engine control. Dr. Tavakoli has served as a reviewer for several journals including IEEE Transactions on Automatic Control, IEEE Transactions on Control Systems Technology, IET Control Theory & Applications, and a number of international conferences . © 2010 JOT© 2010 JOT