Abstract
A sparse substrate integrated waveguide (SIW) slot antenna array and its application on phase scanning are studied in this paper. The genetic algorithm is used to optimize the best arrangement for 8-element and 7-element sparse arrays over an aperture of . Antenna arrays with feeding networks, for steering the main beam pointing to 0° and −15°, are demonstrated with the SIW technology. The comparison between the sparse array and the conventional uniformly spaced array with the same aperture are presented, which suggest that the same beam width can be obtained with the gain decreased by 0.5 or 1 dBi and the number of element reduced by 2 or 3, respectively. The sparse antenna array with beam scanning ability presented in this paper shows that, while the beam scanning in the range of ±15°, the gain fluctuation is less than 0.3 dBi and the side lobe level is lower than −10 dB.
1. Introduction
Phased array antenna has been widely used in modern wireless communication systems for the high gain and agile beam scan [1]. The active phased array has transmitting/receiving (T/R) module with each element, which can improve the performance but will increase the cost of the whole array. Besides, in order to obtain a narrow main beam, a large radiation aperture is needed; the element number of the traditional uniformly distributed phased array will increase. The cost introduced by control components, T/R modules, and power dividers will limit the use of phased arrays. In some applications with the requirement of narrow beam width and high gain, the phased array with reduced number of cells attracts more attention.
Sparse array was studied with fewer elements arranged over the same aperture compared to conventional full feeding uniformly spaced array since the 1960s [2]. Theoretically, the low side lobe level (SLL) can be obtained by optimizing the feeding current and position of each element. Some algorithms are employed to large scale sparse array optimizations, such as genetic algorithm (GA) [3], particle swarm optimization [4], and Harmony Search Algorithm [5]. Among the optimization methods, genetic algorithm is one of the most popular optimization techniques used for side lobe level reduction [6–9], since GAs [10] are well suited for sparse and thinning array optimization. GA is a global and random search algorithm that simulates natural selection and evolution. It searches through the total solution space and can find the optimal global solution in a domain. Although GA was applied for optimizing sparse array widely, few sparse arrays with phased scanning have been reported with demonstration and measurement results.
In this paper, the genetic algorithm is applied for optimizing the position of each element in an equal amplitude feeding sparse array with an aperture of , where is the free space wavelength at the center frequency 14 GHz. The sparse arrays with antenna element number of 7 and 8 are investigated separately, which SLLs are lower than −10 dB at both broadside and −15° direction. Each sparse array consists of antenna elements, phased shifters, and power dividers based on substrate integrated waveguide (SIW).
The paper is organized as follows. The design process of the sparse phased array is presented in Section 2. Measurement results and analysis are given in Section 3, and the conclusion is drawn in Section 4.
2. Design of the Sparse Array
2.1. Genetic Algorithm (GA)
A linear sparse array is studied in this section. The array factor with equal feeding amplitude can be given in [11] by where is the distance between the element and the element 1. is the number of the elements of the array. The program in MATLAB GA toolbox is used to optimize the best arrangement of the linear sparse array. The GA flowchart is shown in Figure 1. is set as optimization variable, and SLL is chosen as fitness function, which is expressed as where is the peak main lobe level.
For the 8-element case, the position of the first element is 0 and the 8th element is . The 8-number vector is coded into a population with 400 randomly generated individuals. In order to avoid the dead loop, as well as considering the speed of the optimization convergence, after 500 iterations, a set of element arrangement will be generated. When the number of element is 7, the situation is the same as the 8-element case except the population is coded by 7-number vectors. The optimized arrangements for both cases are shown in Table 1.
2.2. SIW Slot Antenna
By using the optimized position of the sparse arrays, we construct the 8-element and 7-element sparse antenna arrays which are shown in Figure 2. They are both composed of SIW radiation slots, phase shifters, parallel power dividers, and 50-ohm grounded coplanar waveguide- (GCPW-) to-SIW transitions. The SIW metallic pins are with the radius of 0.25 mm and the distance of 1 mm.
(a)
(b)
(c)
The two radiation slots in each SIW path are antenna arrays, that is, the substrate integrated waveguide slot antenna unit; the detail of the slot array antenna unit is shown in Figure 2(c); the substrate of the SIW is with the thickness of 1.524 mm and with the relative permittivity of 3.5. All the dimensions of the slot antenna array unit are listed in Table 2. The S11 and radiation of the slot array antenna unit are shown in Figure 3. The −10 dB impedance bandwidth is from 13.4 to 15.2 GHz (relative bandwidth is 12%), with a gain of 7.86 dB at 14 GHz. The beam width is 140° in the E-plane, which indicates that it can be used as a unit antenna for a wide-angle coverage phased array.
(a) S11
(b) Radiation pattern at 14 GHz
2.3. Feeding Network
The feeding networks with phase shifters are designed to steer main beam pointing to 0° and −15° for each sparse array. The structure of Y-type-equal SIW power divider is shown in Figure 4. The distance between the two output ports of each power divider dai is different from each other. A metallic pin is placed in the center line to equally deliver the input power to two ports. The distances dyi and dxi can be optimized to reduce the reflection from the SIW branches and bends [12]. And the width of SIW is 7.3 mm to support the TE10 mode in the whole operating frequency band.
The phase of each output port of the parallel power divider is different because of the ununiformly distributed sparse array. In the cases of the main beam steering to 0°, the phase shifters are required for in-phase feeding. For the cases of beam steering to −15°, the phase shifts of the SIW phase shifters are shown in Table 3. The required phase shifts are realized by adjusting the widths and lengths of the phase shifters, which changes the propagation constant and characteristic impedance of the SIW [13]. In the case of phase shift larger than 180°, the radiation slots can be mirrored along the center line of the SIW to provide 180° phase shift in addition. The scanning performance of the array is studied by feeding each element with equal amplitude and gradual phase shifts.
The structure of 50-ohm GCPW-to-SIW transition is shown in Figure 5(a). It is used to transform low impedance SIW to 50 ohm for testing purpose. The metallic pins near GCPW are with the radius of 0.15 mm and distance of 0.6 mm. The 50-ohm GCPW has a width of and a gap of . The simulated parameters are shown in Figure 5(b), which indicate a low reflection coefficient among operation frequency band.
(a)
(b)
After combining the antenna array and feeding network together, four SIW antennas including 8-element and 7-element sparse arrays pointing to 0° and −15° are simulated and fabricated. The measurement results of S11 of the whole sparse array with feeding networks are shown in Figure 6.
3. Measurement and Analysis
Four sparse array antennas with feeding networks are demonstrated by low-cost single-layer printed circuit board (PCB) process, which are shown in Figure 7, each sparse array is composed of top layer and bottom layer. The substrate is with a dielectric permittivity of 3.5, dielectric loss tangent of 0.0018, which thickness is 1.524 mm with a dimension of 12.4 × 14.7 cm2. The simulated and measured radiation patterns of the sparse array antennas are shown in Figure 8. The measured results agree well with the simulated ones, and the SLLs lower than −10 dB and cross polarization lower than −20 dB for all arrays pointing to different angles.
(a) 8-element sparse array
(b) 7-element sparse array
(a)
(b)
(c)
(d)
Furthermore, the comparison between the sparse array and conventional half-wavelength spaced array with 10 elements as well as the phase scanning performance of the sparse array is shown in Figures 9 and 10, respectively. Both the sparse array and the uniform 10-element array are with the same radiation slot array antenna unit. It suggests that the sparse array can generate the same beam width as the fully arranged array while the gain decreased by 0.5 or 1 dBi when the element number reduced by 2 or 3, respectively. Both sparse arrays show a good performance during the full scanning range of ±15° with the gain fluctuation less than 0.3 dBi and SLL lower than −10 dB.
(a)
(b)
4. Conclusion
In this paper, the sparse SIW slot antenna arrays with beam scanning ability are studied. GA is used to optimize the locations of the elements to make SLL lower than −10 dB. The 8-element and 7-element sparse arrays are designed by using the optimized arrangement, and SIW technology is used in the feeding network. In order to verify the beam scanning ability of the sparse arrays, the arrays pointing to −15° are realized by using the designed feeding network. Although the broadside gain of sparse arrays is decayed, the same beam width can be obtained with a reduced element number. After being fed with the equal amplitude and gradual phase shift, the phase scanning performance of two sparse arrays is studied. The designs are demonstrated and tested, which indicates a good sparse array which also has the beam scanning property when the proper phase is given.
Data Availability
The data used to support the findings of this study are included within the article.
Conflicts of Interest
The authors declare that there is no conflict of interests regarding the publication of this paper.
Acknowledgments
This work was supported in part by the National Key Technology Support Program (2015BAK05B01 and 2015BAK05B01-01).
Supplementary Materials
The simulation and measurement data used to support the findings of this study are included within the supplementary information file. (Supplementary Materials)