CPW Center-Fed Single-Layer SIW Slot Antenna Array for Automotive Radars

Transcription

1 1 CPW Center-Fed Single-Laer SIW Slot Antenna Arra for Automotive Radars Junfeng Xu, Member, IEEE, Zhi Ning Chen, Fellow, IEEE, Xianming Qing, Member, IEEE Abstract A compact co-planar waveguide (CPW) center-fed substrate-integrated-waveguide (SIW) slot antenna arra is proposed to achieve narrow H-plane beamwidth and low sidelobe levels for automotive radars such as Across-the-Road (ATR) Radar. The antenna consists of an arra of SIW slot elements, a CPW-SIW transition and a CPW power divider. A new tpe of CPW-SIW transition is proposed to minimie the blockage aperture of the slot antenna arra for suppressing the sidelobe levels of the arra. A parallel center feeding configuration is applied to avoid beam squinting. The antenna prototpe printed onto a single-laer Rogers 588 is with 32 4 slot elements and an overall sie of 195 mm 4 mm.79 mm. Measured results show that the proposed antenna ehibits gain of > 22.8 i, efficienc of > 67%, return loss of > 1, sidelobe level of < -21, and a fied boresight beam of < 4.6 in H-planes over GH, suitable for ATR radar. Inde Terms Co-planar waveguide (CPW), substrate integrated waveguide (SIW), slot antenna arra, sidelobe, fied beam, single-laer, Across-the-Road (ATR) Radar, automotive radar. A I. INTRODUCTION ntenna is one of the ke components in automotive radar applications [1] [3]. Across-the-Road (ATR) radar, a sensor testing the speed of moving vehicles, is such an eample. The specifications of a K-band ATR radar in U. S. are given in [4]. The assigned frequenc band is 24.5 GH to GH, or.8% operating bandwidth. The critical electrical requirements for ATRs include half power beamwidth (HPBW), sidelobe levels (SLLs), and beam direction. Specificall, the HPBWs in the horiontal and vertical planes are required to be less than 6 and 2, respectivel. The SLLs in the horiontal plane should be less than the -2. In addition, the primar radar beam should be centered horiontall and verticall without beam squinting across the bandwidth. Microstrip antennas have well-known advantages such as low cost, low profile, and eas to integrate with active circuits Microstrip antenna arras have thus been reported to appl in automotive radars with good electrical performance [5] [8]. The main concerns of the microstrip antennas come from the undesired radiation and surface wave coupling/loss. To Manuscript received April 4, 214. The work was supported b Agenc for Science, Technolog and Research (A*STAR), Singapore, Terahert Science & Technolog Inter-RI Program under Grant # and Metamaterials Program under Grant # J. F. Xu was with Institute for Infocomm Research, A*STAR, Singapore. J. F. Xu is now with Massachusetts Institute of Technolog (MIT), 77 Massachusetts Ave, Cambridge, MA, USA. Z. N. Chen is with the Institute for Infocomm Research, A*STAR, Singapore. Z. N. Chen is concurrentl with National Universit of Singapore. ( elecn@nus.edu.sg) X. Qing is with the Institute for Infocomm Research, A*STAR, Singapore. alleviate the surface wave mode, single-laer microstrip antennas usuall have to be printed onto thin substrates [8]. With a thinner substrate, however, the microstrip antenna has more sever conductor and dielectric losses, a narrower bandwidth, and less mechanical robustness [8] [9]. Alternativel, substrate integrated waveguide (SIW) [1] [12] or post-wall waveguide [13] fed antennas feature the same advantages of microstrip antennas because the can be fabricated with the same planar printed technolog. Furthermore, as a waveguide-like structure, SIW does not suffer from the unintentional radiation and surface wave loss, which alleviates the limitation of the thin substrate. These merits offer the SIW antennas the possibilit of achieving a high efficienc. A SIW slot antenna with SLLs as low as -3 in both E- and H-planes have been achieved [1]. However, the end-fed antenna arra suffers from beam squinting against operating frequenc, which does not meet the center beam requirement of the ATR radar. To achieve the required unchanged maimum radiation direction across the operating bandwidth, a center-fed antenna arra is preferred [13] [15]. However, it is difficult to realie a conventional single-laer center-fed waveguide-based slot antenna arra to achieve low SLLs in the H-plane because of the large aperture blockage (slot-free area) in the center portion of the antenna arra. Some studies have reported to reduce the aperture blockage effect and lower the SLLs [13] [15]. In [14], the SLL of -9.5 associated with the aperture blockage is improved to b appling a genetic algorithm to control the slot ecitation distribution. In [15], the first SLL is reduced from to -13 using the E- to H-plane cross-junction power dividers. A post-wall waveguide feeding network for a center-fed antenna arra has been reported [13]. The large blockage area results in a high SLL of -7.8 but is suppressed to using a tapered amplitude distribution. Generall, the width of a post-wall waveguide or SIW is wider than that of a microstrip or co-planar waveguide (CPW), which ma be undesired in the design of antenna feeding network. In this paper, a CPW center-fed SIW slot arra antenna is proposed to achieve a low SLL, narrow and fied beam in the H-plane for the 24-GH ATR radar using normal low-cost single-laer printing circuit board (PCB) process. We first present the design of the slot arra and investigate the effects of the blockage area on the SLLs of the antenna. With the motivation of reducing the blockage area for low SLLs, we emplo a compact, parallel CPW center feeding structure to avoid the undesired beam squinting with frequenc. Finall, the measured results of the antenna arra are discussed and compared with the simulation using CST Microwave Studio [16].

2 2 d h Metal ground d e Blockage or slot-free area 195 mm Radiating slots Vias 4 mm Input port Vias CPW feeding network Metal ground Fig. 1. Top and bottom view of the CPW fed SIW slot arra antenna. II. ANTENNA STRUCTURE AND DESIGN A. Antenna Configuration Fig. 1 shows the top view of the SIW slot arra antenna. There are totall 4 32 slots on the broadwall of the SIWs. In each row, two 16-element linear arras are placed end to end with a distance of between the two starting slots. The central portion of the antenna is occupied b the feeding structure and is thus slot free, which is the bottleneck of sidelobe suppression. The effects of on radiation pattern will be discussed in net subsection. The spacing between the adjacent slots in the H-plane, d h, is λ g /2 (λ g is the guided-wavelength of the SIW). And the spacing between the adjacent slots in the E-plane, d e, is λ w (λ w is the guided-wavelength of the CPW) so that the adjacent linear arras are fed in phase. The detailed geometrical dimensions of the antenna are tabulated in Table I, in which i(i=1 16) and l i(i=1 16) are the offsets and lengths of the slots. Fig. 1 shows the bottom view of the overall slot arra antenna. The input of the arra is connected to an eternal commercial available mini-smp connector. The compact CPW feeding network is located in the center of the antenna with an eight-wa parallel feeding configuration. The antenna is printed onto a single-laer PCB of Rogers 588 with ε r = 2.2 and tanδ =.9 at 1 GH. The thickness of the substrate is t =.79 mm. The conductor used for metalliation is copper with a conductivit of S/m and a thickness of.2 mm. TABLE I GEOMETRY OF THE SLOT ARRAY (unit: mm) l l l l l l l l l l l l l l l l d e 8.7 d h 5.8 B. Design and Analsis of Slot Arra The design of the slot arra follows the method in [1] and is briefl described here. First, the parameter is etracted for a single slot on the SIW with various offsets. Resonant length, resonant conductance, and admittance are obtained for the arra snthesis. Net, Elliott s iterative procedure for a waveguide-fed slot arra [17], including all mutual couplings, is applied for the SIW-fed linear arra to calculate initial slot parameters for a targeted amplitude distribution. Further fine tuning b electromagnetic (EM) simulation finalies the slot parameters for desired SLLs. In order to achieve the requirement of -2 SLL in the H-plane, -26 Talor distribution is chosen. To achieve the <6 HPBW in the H-plane, 32 slots are used. In the simulation of the 32-element linear arra as shown in Fig. 2, two ports ecite each half of the slot arras

3 Amplitude distribution Offset (mm) Length (mm) 3 simultaneousl. The spacing between slots 16 an7,, is first set to be λ g /2 = 5.8 mm. In this case, there is no blockage because equals to the spacing of all other adjacent slots, d h. The distance between the ends of slots 16 an7 is d s = 1 mm. The geometric parameters of the slot arra are listed in Table I. The offsets and lengths of the slots are also shown in Fig. 3. As the slot number decreases, the lengths become smaller and slots are located closer to the center of SIW so that the radiation from the edge slots weakens. The offsets of slots 1 4 are set a little larger than that of slot 5 for a 2- better SLL. To other slots d s Port1 Port2 To other slots Fig. 2. model of the two 1 16 linear arras located end to end with different Fig. 4 compared the simulated amplitude distribution and the theoretical -26 Talor distribution of the slots. The simulated result including all mutual coupling agrees reasonabl with the theoretical one. The simulated H-plane radiation pattern is shown in Fig. 5 ( = 5.8 mm). The -26 peak sidelobe locates at the first one. The SLLs decrease as θ increases. In this case, unfortunatel, d s of 1 mm is too small to accommodate a feeding network and necessar space between the slot and the feeding network to avoid unwanted EM coupling which affects the SLLs [15]. Therefore, there is a tradeoff between the blockage area and the requirement of realiing the feeding network. Fig. 5 also shows the H-plane radiation patterns of the linear arra with different at 24 GH. As increases from λ g /2 (5.8 mm) to λ g (11.6 mm), the innermost SLL changes slightl but the grating lobes in the range of 3 < θ < 75 increase due to the enlarged blockage area. The peak SLL is When becomes 16 mm, the peak and first SLL degrades to -19. To meet the requirement of -2- SLL, the case of SLL is selected with 4.5 margin considering fabrication tolerance. This case with of 11.6 mm results in d s = 6.84 mm, possible to accommodate a compact feeding network and the space to avoid undesired coupling. -2 =5.8 mm =11.6 mm =16. mm Fig. 3. Fig Slot number Slot offsets and lengths of the antenna arra Calculation Slot number Amplitude distribution of the linear arra Fig. 5. H-plane radiation patterns of the linear arras with different at 24 GH. After aligning multiple linear arras side-b-side to form the SIW planar arra, the H-plane SLL of the whole planar arra is almost unchanged. The reason is that SIW is a low-profile waveguide with about 8:1 width-to-height ratio so the TE 2 -mode internal mutual couplings inside the SIW are dominant among all kinds of internal and eternal mutual couplings [1]. After forming the SIW planar arra, onl eternal mutual couplings among slots at different branches of SIW are introduced. The effects of the additional eternal mutual couplings are much smaller than those mutual couplings caused b the TE 2 mode which have alread been included in the SIW linear arra design. It is known that direct tuning the hundreds of parameter of the SIW planar arra in EM simulation is ver time-consuming and inefficient. In the above wa, one can significantl simplif the EM design process b onl optimiing one row of the planar arra with much less parameters.

4 4 C. Feeding Network Design To achieve relativel smaller blockage for low SLLs, CPW, instead of a wide SIW or post-wall waveguide [15], is utilied to configure the power divider to feed the arra antenna. The first step is to select the most compact CPW-SIW transition. Several CPW-SIW transition designs have been reported and four tpical ones are shown in Fig. 6 as the candidates for the compact feeding structure of this design [18] [21]. The transition based on a dipole slot in Fig. 6 [18] and transition with a current probe in Fig. 6 use a λg/4 GCPW stub for impedance matching [19]. Another transition in Fig. 6(c) has a >λ g /4 cavit resonator under the dipole slot [2]. Because all the λ g /4 structures are inside the SIW, it is impossible to move them from the limited H-plane space (-direction) to E-plane space (-direction). The transition depicted in Fig. 6(d) [21] is preferred wherein the dipole slot is net to the short-circuit end of the SIW so that the most compact configuration is realied. In that design, however, a λ w /4 CPW impedance transformer is net to the SIW which cannot be applied directl to form an H-plane (-direction) compact design. The impedance transformer outside the SIW can be removed along -direction and the impedance matching network can be designed along -direction because of more space there. The removal of the impedance transformer at each stage leads to an inter-stage impedance other than 5 Ω in the feeding network. In this wa, =11.6 mm is achieved as shown in Fig. 7. Distance d 3 = 2 mm is set to avoid the unwanted EM coupling between the transition and the slots which affects the slot ecitation and SLLs [15]. Width of w 2 = 1.9 mm sets a limit of the widest CPW width used in the power divider. If the two λ w /4 CPW impedance transformers were positioned along the -direction [21], would increase to 16 mm. The radiation pattern of this arra configuration is shown in Fig. 5, where = 16 mm leads to a higher SLL up to -19. λ g /4 λ g /4 Fig. 7. d 3 w 2 To net stage and impedance transformer First stage of the feeding network. Design of the feeding network is illustrated in Fig. 8. The E-plane spacing (-direction) between the adjacent SIW is set to be the guided wavelength of CPW at 24 GH, λ w, so that the phase of S 21 equals to the phase of S 31. Because of the structure smmetr, this configuration guarantees the phase balance at the eight outputs. The input impedance of the half-wave slot dipole and the CPW with a length of l c determine Z 1 as shown in Fig. 8. If the characteristic impedance of the CPW is low, Z 1 is small and Z 2 (roughl Z 1 /4) is thus even smaller. This small Z 2 requires an impedance transformer with a width eceeding the pre-assigned w 2 of 1.9 mm. Here, the characteristic impedance of the CPW is set to be 83 Ω. With the dimensions in Fig. 9, the simulated Z 1 is (211 - j3) Ω. The imaginar part of Z 1 is not necessaril ero and can be canceled out at the net stage. The simulated Z 2 = (5 + j37) Ω is not eactl Z 1 /4 because of the effects of the junctions, taken into account onl in simulation. To match Z 2 to 1 Ω, a 46 Ω CPW line with an electrical length of 3 is needed. Because the width of the 46 Ω CPW line is 1.5 mm, it leads to a wide w 2 greater than pre-assigne.9 mm. Instead, we used a 55-Ω CPW line, with an electrical length of 38, which is onl.8 mm wide, for the impedance transformer. Although in this case Z 2 is matched to Z 3 = (11 j4) Ω, the overall impedance matching is still acceptable as shown later. The rest of the feeding network includes two parallel 1 Ω CPW lines connecting to the 5Ω input line. Fig. 9 shows the detailed geometr of the rest of the feeding network. >λ g /4 λ w /4 (c) (d) Fig. 6. Four tpes of CPW-SIW transition in [18] [21].

5 Phase (degree) Amplitude () S 11 () 5 1Ω Port2 Z 3 55Ω, 38º λ w Port3 Z , and -9.65, respectivel. In Fig. 11, the phase of S 21, S 31, S 41, and S 51 are 173, 171, 171, an69, respectivel. Therefore, good impedance matching and amplitude/phase balance are achieved b the proposed feeding network S 11 S 21 S 41 S 31 S Fig. 8. 5Ω input Port1 Port4 Z 2 Port5 Input impedances at various reference planes of the feeding network. Port2 w SIW Frequenc (GH) Fig. 1. Simulated amplitude of S-parameters of the CPW feeding network. l t l s w v l c Z 1 d v Port1 s g Frequenc (GH) S 21 S 41 S 31 S 51 Fig. 11. Simulated phase of S-parameters of the CPW feeding network. t 2 t 1 g 2 s 2 p v s 3 III. EXPERIMENTAL RESULTS The antenna was prototped and measured in a full anechoic chamber. The antenna prototpe shown in Fig. 12 is with an overall sie of 195 mm 4 mm.79 mm. s 1 Fig. 9. Detailed geometr of the feeding network (unit: mm). CPW-SIW transition, s =.2, g=.1, w v = 1, l t =.5, l s = 4.3, w SIW = 6.2, l c =.95, impedance transformer part. s 1 = 1.2, s 2=.3, s 3 =.8, g 2 =.25, t 1 = 2.9, t 2 = 1.25, p v =.8, d v =.4. The simulated amplitude of the S-parameters of the feeding network is shown in Fig. 1. Onl four output ports, Ports 2 5 are shown because of the smmetrical structure. At 24 GH, S 11 is -18. S 21, S 31, S 41, and S 51 are -9.5, -9.53, Fig. 12. Photo of the antenna arra: top view, bottom view.

6 Boresight gain (i) Half power beamwidth (degree) S 11 () Sidelobe level () 6 Fig. 13 shows the measured and simulated S 11 of the antenna arra prototpe. Reasonable agreement has been achieved between the measurement and simulation. The simulated S 11 is less than in GH and the measured S 11 is less than in GH. upward frequenc shift is also observed between measurement and simulation. Compared to the simulated SLL of the linear arra at 24 GH in Fig. 5, the simulated SLL of the planar arra without an tuning of the slots is The SLL of the planar arra is almost unchanged compared to the linear arra Frequenc (GH) Fig. 13. Measured and simulated S 11 of the antenna arra. Fig. 14 shows the measured and simulated boresight gain of the antenna arra. In the measurement, a mini-smp adapter was used [22]. In Fig. 14, the measured gain includes the insertion loss of such an adapter. Compared to the original simulated gain, the modified simulated gain shifting about.1 GH upwards agrees better with the measured one. The difference between measured and shifted simulated gain attributes to the actual antenna loss which ma be caused b loss dielectric and higher than that in simulation. In GH, the measured gain is higher than 23 i and the maimum gain is 24 i at GH Frequenc (GH) Fig. 15. Measured and simulated H-plane sidelobe levels of the antenna arra. Fig. 16 shows the measured and simulated H-plane HPBW of the antenna arra. Over the range of GH, the simulated H-plane HPBWs keep less than 4.6. In GH, the measured H-plane HPBWs are less than 4.6. In the E-plane, the measurement shows that the HPBWs are less than 2 across the band of GH (Shift) (Original) Frequenc (GH) Frequenc (GH) Fig. 14. Measured and simulated boresight gain of the antenna arra. Fig. 15 shows the measured and simulated SLLs of the antenna arra. In GH, the simulated SLL is lower than and the lowest SLL is at 23.9 GH. The measured SLL is lower than -21 in GH. An Fig. 16. Measured and simulated H-plane HPBW of the antenna arra. Fig. 17 shows the measured and simulated efficienc of the antenna arra. The antenna efficienc takes both radiation efficienc and mismatch loss into account [23]. The simulated efficienc is higher than 7% in GH. Because of the agreement between the measured and simulated beamwidths, the simulated directivit (shifting.1 GH upwards) and measured gain (including the insertion loss of the mini-smp adapter) are used for the calculation of measured efficienc. The measured efficienc is higher than 67% within the range of

7 Efficienc (%) GH. The high antenna efficienc attributes to the compact CPW feeding network with minimies the path loss and the low-loss propert of SIW. It is observed that the measured efficienc is shifted upwards compared to simulation as well Frequenc (GH) Fig. 17. Measured and simulated antenna efficienc. From Figs , it is observed that all the measured responses shifted upwards.1 GH compared with the simulation. Therefore, in the comparison between the measured and simulated radiation patterns, we consider this frequenc shift. Figs show the comparison of the measured and simulated radiation patterns of the antenna arra in the H- and E-planes, respectivel, with a +.1-GH shift. In addition, the main beam of the antenna keeps at boresight without an squinting as epected. For the H-plane radiation patterns over the band from 24.5 GH to GH in Fig. 18, the measured beamwidths agree well with the simulated ones. The measured inner-most SLLs increase because the are sensitive to fabrication tolerance and errors. In Fig. 19, the measured beamwidths and peak SLLs in the E-plane are also ver close to the simulated ones. However, the first sidelobes are not smmetrical in the measurement. Because of the E-plane smmetr in antenna configuration, the radiation pattern should be smmetrical as the simulation shows. Thus, the possible asmmetr in measurement and/or fabrication tolerance ma lead to the asmmetr of the measured results (c) Fig. 18. Measured and simulated H-plane radiation patterns of the antenna arra. at 24.5 GH and simulation at GH, measurement at GH and simulation at 24.5 GH, (c) measurement at GH and simulation at GH.

8 (c) Fig. 19. Measured and simulated E-plane radiation patterns of the antenna arra. at 24.5 GH and simulation at GH, measurement at GH and simulation at 24.5 GH, (c) measurement at GH and simulation at GH. The measured results shown in Figs suggest that, the workable frequenc range of the antenna is GH, the same as the officiall assigned band of the ATR radar [4]. In this band, all the requirements of the ATR radar on HPBWs, SLLs, beam directions are well met. Furthermore, the measured gain, efficienc, and return loss are higher than 22.8 i, 67%, an, respectivel. IV. CONCLUSION In automotive radars, it is a required for an antenna arra to achieve low SLL and narrow H-plane HPBW without an beam squinting. The proposed CPW center-fed planar SIW slot arra antenna has been validated to be with smaller blockage. The proposed antenna has eperimentall demonstrated high gain greater than 22.8 i, HPBW narrower than 4.6º, low SLL less than -21 and stable boresight radiation. The compact antenna printed onto a single-laer PCB has been with a simple structure and low fabrication cost, which is suitable for ATR radars operating at 24 GH. REFERENCES [1] W. Menel and A. Moebius, Antenna concepts for millimeter-wave automotive radar sensors, Proceedings of the IEEE, vol. 1, no. 7, pp , 212. [2] T. Hong and S. Chung, 24 GH active retrodirective antenna arra, Electron. Lett., vol. 35, no. 21, pp , [3] S. Shi, J. Purden, J. Jin, and R. A. York, A 24 GH wafer scale electronicall scanned antenna using BST phase shifters for collision avoidance sstems, in IEEE Int. Smp. Antennas Propag., 25, pp [4] Speed-Measuring Device Performance Specifications: Across-the-Road Radar Module, U.S. Department of Transportation, [Online]. Available: ages/speedmeasuringdeviceperform.pdf [5] H. Iiuka, T. Watanabe, K. Sato, and K. Nisikawa, "Millimeter-wave microstrip Arra antenna for automotive radars," IEICE Trans. Commun., vol. E86-B, no. 9, pp , Sep. 23. [6] M. Slovic, B. Jokanovic, and B. Kolundija, High efficienc patch antenna for 24 GH anticollision radar, in International Conference on Telecommunications in Modern Satellite, Cable and Broadcasting Services, vol. 1, pp. 2 23, 25. [7] K. Sakakibara, S. Sugawa, N. Kikuma, and H. Hiraama, Millimeter-wave microstrip arra antenna with matching-circuit-integratedradiating-elements for travelling-wave ecitation, in European Conference on Antennas and Propagation, pp. 1 5, 21. [8] L. Han and K. Wu, 24-GH bandwidth-enhanced microstrip arra printed on a single-laer electricall-thin substrate for automotive applications, IEEE Trans. Antennas Propagat., vol. 6, no. 5, pp , 212. [9] D. Poar and D. Schaubert, Microstrip Antennas: The Analsis and Design of Microstrip Antennas and Arras. John Wile & Sons, [1] J. F. Xu, W. Hong, P. Chen, and K. Wu, Design and implementation of low sidelobe substrate integrated waveguide longitudinal slot arra antennas, IET Microw., Antennas Propagat., vol. 3, no. 5, pp: , 29. [11] J. F. Xu, Z. N. Chen, X. Qing, and W. Hong, Bandwidth enhancement for a 6 GH substrate integrated waveguide fed cavit arra antenna on LTCC, IEEE Trans. Antennas Propagat., vol. 59, no. 3, pp , 211. [12] P. Chen, W. Hong, Z. Kuai, J. Xu, H. Wang, J. Chen, H. Tang, J. Zhou, and K. Wu, A multibeam antenna based on substrate integrated waveguide technolog for MIMO wireless communications, IEEE Trans. Antennas Propagat., vol. 57, no. 6, pp , 29. [13] K. Hashimoto, J. Hirokawa, and M. Ando, A post-wall saveguide center-feed parallel plate slot arra antenna in the millimeter-wave band, IEEE Trans. Antennas Propagat., vol. 58, no. 11, pp , 21.

9 [14] P. Sehun, Y. Tsunemitsu, J. Hirokawa, and M. Ando, Center feed single laer slotted waveguide arra, IEEE Trans. Antennas Propagat., vol. 54, no. 5, pp , 26. [15] Y. Tsunemitsu, S. Matsumoto, Y. Kaama, J. Hirokawa, and M. Ando, Reduction of aperture blockage in the center-feed alternating-phase fed single-laer slotted waveguide arra antenna b E- to H-plane cross-junction power dividers, IEEE Trans. Antennas Propagat., vol. 56, no. 6, pp , 28. [16] CST Microwave Studio, Computer Technolog AG, [Online]. Available: [17] R. S. Elliott andw. R. O Loughlin, The design of slot arras including internal mutual coupling, IEEE Trans. Antennas Propagat., vol. AP-34,pp , Sept [18] D. Deslandes, and K. Wu, Integrated transition of coplanar to rectangular waveguides, in International Microwave Smposium, USA, pp , 21. [19] D. Deslandes, and K. Wu, Analsis and design of current probe transition from rounded coplanar to substrate integrated rectangular waveguides, IEEE Trans. Microwave Theor and Tech., vol. 53, no. 8, pp , Aug. 25. [2] A. Patrovsk, M. Daigle, and K. Wu, Millimeter-wave wideband transition from CPW to substrate integrated waveguide on electricall thick high-permittivit substrates, in European Microwave Conference, pp , 27. [21] S. Lin, S. Yang, and A. E. Fath, Development of a novel UWB Vivaldi antenna arra using SIW technolog, Progress In Electromagnetics Research, PIER 9, , 29. [22] Mini-SMP adapter datasheet, Rosenberger, [Online]. Available: [23] C. A. Balanis, Antenna Theor: Analsis and Design. John Wile & Sons, pp , 25. 9