A WLAN Dual-Polarized Beam-Reconfigurable Antenna

A novel wireless local area network (WLAN) antenna with dual-polarized and beam-reconfigurable ability is developed for WLAN 2.4 GHz band application. The antenna is composed of a horizontally polarized (HP) antenna and a vertically polarized (VP) antenna. The radiation pattern of the HP antenna can be switched between an omnidirectional beam and four directional beams in the azimuth plane, while the VP antenna can generate an omnidirectional beam and eight directional beams. Then the antenna prototype is fabricated and measured. The directional beam gain of the HP antenna varies from 2.7 to 3 dBi, and that of the VP antenna varies from 2.5 to 3.5 dBi. The maximum difference between the measured and simulated directional beams is 0.9 dBi. The measured and simulated reflection coefficients of the antenna are below −10 dB for all the beam states, and the isolation between the two polarizations is less than −18 dB in the WLAN 2.4 GHz band. The measured results agree well with the simulated ones, showing a good application prospect in WLAN system in the future.


I. INTRODUCTION
A S INDUSTRY and people's lives demand data traf- fic [1], the 802.11protocol suite has been continuously optimized and developed.Wireless local area network (WLAN) have been put into use in many public places such as schools, theaters, stations and airports.As the hub responsible for receiving and sending data in WLAN, the antenna plays a vital role.In the past, WLAN antennas could operate in multiple frequency bands to maximize the use of spectrum resources [2], [3], [4], [5], [6], [7], [8].However, with the continuous increase of WLAN users, electromagnetic signal coverage is limited, electromagnetic waves are blocked by obstacles, and link crosstalk problems in dense networking environments are becoming more and more serious.To solve these problems, dual polarization (DP) [9] and pattern reconfigurable [10], [11] techniques can be introduced to improve the performance of the WLAN system.DP antennas have twice as much communication capacity as the single polarization antennas when good polarization isolation is realized.Pattern reconfigurability technique can provide multiple communication paths for users and network access points, which can effectively improve the electromagnetic signal coverage.
Currently most of the beam-scanning antennas are implemented based on phased arrays, and the beam scanning is realized by changing the feeding phase by using phase shifters.The implementation for DP radiation usually includes two categories: designing orthogonally placed feeding structure under the same radiator [12], [13], [14]; integrating two kinds of polarized radiators and feeding each polarized radiator separately [15], [16], [17].Although the antennas above can realize the DP beam scanning, they are not quite suitable for WLAN systems due to high cost and large size.
Various WLAN antennas are currently available.Traditional WLAN antennas only work in the WLAN 2.4 GHz or 5 GHz frequency band, and can provide an omnidirectional beam [18], [19], [20].Later, some WLAN antennas can also achieve frequency reconfiguration feature [21], [22], [23].In order to improve the signal coverage, WLAN antennas with beam switching function have been investigated, which can produce multiple high-gain directional beams [24], [25], [26].Then, in order to further improve the anti-interference performance, on the basis of orthogonal placement of radiators, a DP WLAN antenna emerges as required [9].Nevertheless, the antennas mentioned above cannot produce DP radiation with beam switching characteristic among multi-directional beams and an omnidirectional beam.
It is well known that DP radiations can increase the system capacity under the premise of achieving good polarization isolation.The omnidirectional beam can cover the entire space at the same time, which is suitable for the scene with fewer users and great mobility.The directional beams with high gain have advantages of penetrating obstacles such as floors and walls.Even if one beam cannot pass through obstacles, other directional beams can be used for reflection, diffraction and other multi-path ways to overcome fading and produce a better coverage.Consequently, the above antennas cannot effectively solve the problems of link crosstalk, obstacle occlusion and limited coverage faced by current WLANs.
Therefore, in this article, a novel WLAN antenna with DP and beam-reconfigurable abilities is presented for WLAN 2.4 GHz band application.The antenna can produce HP radiation by a horizontally placed Alford loop antenna, and beam switching is achieved by controlling the surface current at both ends of the rectangular slots etched on the ground through the on-off state switching of the diodes.Therefore, the HP antenna can be switched flexibly between the omnidirectional-beam and the four directional-beam states.On the other hand, the VP radiation is generated by the vertically placed monopole, and the beam reconfiguration is realized by controlling the four strip reflectors distributed around the monopole.Finally, the VP antenna can be switched among an omnidirectional-beam state and eight directional-beam states.In addition, a choke plate with symmetrical L-slot is designed, which can effectively suppress the secondary radiation of induced current on the outer conductor of the coaxial cable for HP antenna.The DP beam reconfigurable antenna proposed in this paper can effectively improve the capacity and anti-interference performance of WLAN system, and its switchable high-gain directional beams can cover a larger region, and show a better adaptability in communications in complex environment.

II. ANTENNA DESIGN AND ANALYSIS A. ANTENNA CONFIGURATION
The geometry of the proposed WLAN antenna is shown in Fig. 1.The antenna is divided into three parts: HP antenna, VP antenna and choke plate.The components of the antenna are manufactured by printed circuit board (PCB) technology, and the dielectric boards are FR4 (ε r = 4.2, tan δ = 0.025).After all the components are processed, the pins and slots are  used to assemble these PCBs.Fig. 2 (a) depicts the layers of the HP antenna.The HP antenna board is consisting of substrate 1 and substrate 2, and the thickness of substrate is 1.6 mm.There are three metal layers, namely the top layer, the middle layer, and the bottom layer.The three metal layers are all symmetrical structures, and the top and middle layers are rotated 20 • counterclockwise relative to the coordinate axis.The inner and outer conductors of cable_HP are soldered to the top and middle layers, respectively.Fig. 2 (b) shows the top layer of HP antenna.It uses the Alford loop antenna as the radiator, and the radiator is fed by a 1 to 4 microstrip power splitter.The bias circuit provides the bias voltage for the capacitor and diode.The PIN diodes on the top layer are BAR50-02V.In the simulation, the equivalent circuit of the diode is a resistance of 4 when it is forward biased, and a resistance of 3000 and a capacitance of 800 pF in parallel when it is not biased.The capacitors mounted on the top layer are GRM1885C with a capacitance of 1 pF for DC blocking.In addition, in order to avoid the interference of the large electric length of the bias line on the pattern, the bias line is divided into several small segments less than a quarter wavelength in length, and each of which is connected by inductance (81-LQW15AW, 75 nH).The positive electrode of the diode and one end of the capacitor are welded together on a pad, and the negative electrode of the diode and the other end of the capacitor are welded on the other two pads, and then connected to the circular ground on the middle layer through the posts, as shown in the upper right corner of Fig. 2 (b).Fig. 2 (c) shows the circular ground of the middle layer.The circular ground is etched with four rotationally symmetrical rectangular slots spaced 90 • apart, and it's controlled by the diodes on the top layer that can be used as a reflector to achieve the beam reconfiguration function.As shown in Fig. 2 (d), the bottom layer is a isolation structure, and it is composed of 20 fan-shaped units.This isolation structure can improve the isolation performance between DP waves, as shown in Fig. 8. Fig. 3(a) shows the composition of the VP antenna.The VP antenna is composed of substrate3, substrate4 and substrate5, wherein substrate3 and substrate4 are placed orthogonal and then vertically inserted into the slots of the substrate5.Finally, the three parts of substrate3, substrate4, and substrate5 are connected by pins.The thickness of the substrates used in the VP antenna is 1 mm.Fig. 3 (b) shows the front view of substrate3.Two metal strips placed symmetrically can act as reflectors.When the diode at the bottom of one strip is turned on, the electrical length of the strip becomes L r1 + L r2 , and a directional beam can be produced in the opposite direction.In addition, two groups of symmetrical bias structures are placed outside in Fig. 3 (b), each group has three bias lines.The outermost two bias lines are connected with the diodes and the capacitor of the HP antenna through pins, and the inner one bias line is connected with the diodes on the metal strip of VP antenna.The back of substrate3 has no metal structure.As shown in Fig. 3 (c), the radiator of the VP antenna is a vertically inverted triangular monopole antenna with an inverted U-shaped slot etched in the middle of substrate4.This monopole can generate an omnidirectional beam.Fig. 3 (d) shows the back view of substrate4.The rectangular metal patch1 is located on the back of substarte4.The inner conductor of the coaxial cable_VP is connected to patch1, and the outer conductor is connected to the metal ground on substrate5, as shown in Fig. 3 (e).In this way, the vertically inverted triangular monopole antenna can be fed by cable_VP.

B. PATTERN RECONFIGURATION MECHANISM
The beam reconfiguration function of the HP antenna is realized by changing the current distribution of the circular ground.Fig. 4 (a) shows the current distribution when the antenna operates in a directional beam state at 2.44 GHz.At this time, diodes D1 and D2 are in the on state, while D3 and D4 are in the off state.The surface current distributed on the rectangular slot with D1 and D2 can pass smoothly through the diode, and hence the slot is equivalent to be shorted at this moment.Consequently, this part of circular ground has an electrically large size, and it can act as a reflector.Conversely, the current cannot pass through D3 and D4, so it is cut by the slots.Finally, this directional beam radiates toward ϕ = 20 • (recorded as HP1 state).Therefore, by sequentially controlling the on-off state of adjacent diodes, three other directional beams pointing to ϕ = 110 • , ϕ = 200 • , and ϕ = 290 • can be provided (recorded as HP2 ∼ HP4 states).Similarly, when all diodes are off, the current on the whole circular ground plate is not continuous, as shown in Fig. 4 (b).Thus it cannot be used as a reflector at this time, so the omnidirectional beam can be provided (recorded as HP5 state).The reconfigurable beams and corresponding states of diodes of the HP antenna are shown in Table 1.
The VP antenna uses a vertically placed inverted triangular monopole antenna as a radiator, and four metal strips are   also placed vertically around the monopole antenna.Then, a diode is welded to the bottom of each metal strip, and the metal strip acts as a reflector when the diode is switched on.Fig. 5(a) shows the arrangement of strips and diodes of the VP antenna.As shown in Fig. 5 (b), when the diode D5 is on, the current can pass through the diode smoothly.At this time, the electrical size of the metal strip becomes L r1 + L r2 , and the electrical size becomes larger equivalent to a reflector.At this time, the diode D7 is off, so the current is concentrated only on the L r1 part.Similarly, diodes D6 and D8 are also in off state, so the corresponding strips cannot act as reflectors.Therefore, a directional beam pointing to ϕ = 0 • is produced.Based on the above mechanism, different directional beams and omnidirectional beam can be generated by controlling the on/off state of the diodes.As shown in Table 2, the VP antenna has eight directional beams (VP1 ∼ VP8) and one omnidirectional beam (VP9) in total.

C. CHOKE PLATE AND ISOLATION STRUCTURE
As can be seen from Fig. 1, cable_HP passes vertically from the HP antenna through the VP antenna substrate5.So the induced current on the outer conductor of cable_HP would generate a secondary radiation and hence distort the radiation   pattern.Taking beam VP1 as an example, after adopting cable_HP, the beam direction has changed from ϕ = 0 • to ϕ = 30 • , and there is also a large backward radiation, as shown in Fig. 6 (the black dotted line and blue dotted line).Therefore, a choke plate structure with a symmetrical L-slot etched is introduced, as shown in Fig. 7 (a).The L-shaped slot can be considered as an integrated quarter-wavelength choke, which can suppress surface current on the cable.The structure operates as a pair of LC resonators in parallel, then in series with an inductor, and the equivalent circuit of this structure is shown in Fig. 7 (b).The values of inductor L1,   L2, L3 and capacitor C1, C2 are related to the dimensions of the L-shaped slot, and also determine the working frequency band of the choke plate.The final optimized dimensions are given in Fig. 6.Finally, the radiation pattern of VP1 after introducing the choke plate is basically consistent with the initial state (black dashed line in Fig. 6).
Fig. 2(d) shows an isolation structure that enhances the isolation between HP and VP radiations.After other structures are completed, the isolation parameter is compared to present the effectiveness of the isolation structure.In the simulation, the isolation structure is removed, and other structures are exactly the same.In the process of collecting the isolation parameter, HP antenna works in the HP1 state,   and the VP antenna switches from VP1 to VP9 states in order.The final isolation results are shown in Fig. 8, and it can be seen that the antenna isolation parameter is lower than −15 dB in the range of 2 ∼ 3 GHz when no isolation structure is introduced.After the isolation structure is adopted, the isolation can be improved to be less than −23 dB over the same band.

III. RESULTS AND DISCUSSION
The proposed design is verified by fabricating and measuring the antenna.The antenna prototype is shown in Fig. 9.As can be seen from the photographs, the antenna prototype is consisting of five PCBs through pins, and the cables_HP  and cable_VP are used to feed the HP and VP antennas, respectively.Fig. 10 shows the isolation between the two polarizations, which is obtained by switching the states of the VP antenna from VP1 ∼ VP9, when the HP antenna remains in the state of HP1.The results show that the simulated isolation between the VP and HP antenna is less than −23 dB in the band of 2 ∼ 3 GHz, and the measured isolation is less than −18 dB.The measured result is slightly worse than the simulation result, especially when the VP antenna is operating in the directional beam states (VP1 ∼ VP8).This may be due to assembly errors.However, it can still meet the requirements of WLAN systems.
Fig. 11 shows the reflection coefficients of HP antenna in varied operation states.In the band from 2.4 to 2.48 GHz, the simulation and measurement results are less than −10 dB.It should be noted that the working state of HP5 is different from that of HP1 ∼ HP4, because when it works in the omnidirectional radiation state (HP5), the diodes D1 ∼ D4 are all in the off state, and the current distribution on the middle layer is different from the current distribution in the directional radiation state.Fig. 12 shows the reflection coefficient results of VP antenna.It can be seen that the measured results are in good agreement with the simulated results, which are lower than −10 dB in the band from 2.4 to 2.48 GHz.
Taking the HP1 state of HP antenna and the VP2 state of the VP antenna as examples, Fig. 13 (a     to an unideal soldering of choke inductors on the bias line connected to diode D2, leading to effects on the radiation.Overall, the simulated results are in general agreement with the measured results.Fig. 16 shows the azimuth-plane patterns in HP1 ∼ HP4 states, and the measured azimuthplanes of these four patterns are ϕ = 20 • (HP1), ϕ = 110 • (HP2), ϕ = 230 • (HP3), and ϕ = 290 • (HP4), respectively.The angle of maximum gain is about θ = 65 • , and the simulated and measured result are basically consistent with each other.Therefore, it can be inferred that the HP antenna can generate four directional beams and one omnidirectional beam.
Similarly, the horizontal plane and vertical plane (azimuth plane) patterns were also measured for each operating state of VP antenna.Fig. 17 shows the horizontal plane patterns in VP1 ∼ VP9 states (at θ = 90 • ).Among them, the measured results of the VP1, VP3 and VP9 are shifted compared with the simulation results.This may be a result of errors in the dimensions of the slots on the substrate5 during machining, causing the choke plate to be positioned inaccurately during assembling, and hence the choke plate may not work in an optimal condition.And the measured beam width of VP6 is narrower than the simulated result, which may be caused by the tilting of substrate4 during assembling.The simulation results of VP1 ∼ VP9 are basically consistent with the measurement results.In addition, the azimuthplane patterns of VP antenna are also measured, and the measured angles are ϕ = 0 and ϕ = 315 • , respectively.Fig. 18 shows the azimuth-plane patterns of VP antenna.Compared with the simulation results, the angle of maximum gain is about θ = 45 • .In addition to the reasons analyzed in Fig. 17, it is also possible that an unideal soldering of choke inductors has caused some effects of the bias lines on the antenna radiation.In summary, the VP antenna can generate eight directional beams and one omnidirectional beam in the azimuth plane.
The realized gain of the directional beams of the two polarizations in varied working states (HP1 ∼ HP4, VP1 ∼ VP8) are given in Fig. 19.The simulated realized gain of the HP antenna fluctuates around 3 ∼ 3.5 dBi, which is slightly greater than the measured result (the difference is less than 0.4 dBi).For the VP polarization, the simulated realized gain varies from 2.9 to 3.9 dBi, with a maximum difference of 0.9 dBi (VP4) greater than that in measurement.
A comparison of the antenna proposed in this paper with previous related works is shown in Table 3.Compared with the WLAN antennas proposed in [18], [21], [25], the antenna proposed in this paper not only achieves beam reconfiguration, but also has DP performance, and hence the communication capacity is higher.Compared with [9], the proposed antenna adopts fewer feed ports to effectively reduce the system complexity, while the DP and beam reconfigurable function are achieved.

IV. CONCLUSION
In this article, a novel WLAN DP reconfigurable antenna is developed.The frequency band of the antenna can cover 2.4 ∼ 2.48 GHz.This antenna can provide five types of HP beams and nine types of VP beams.All beams in varied polarization states can be flexibly switched.Finally, the measurement results show that the isolation between two polarizations is greater than 18 dB, and the maximum gain of HP antenna is 3 dBi and that of VP antenna is 3.5 dBi.The simulated and measured results are basically consistent.This DP antenna has great potential to be used in WLAN systems and to increase the communication capacity.

FIGURE 1 .
FIGURE 1. Geometry of the developed WLAN antenna.

FIGURE 5 .
FIGURE 5. VP antenna.(a) Diodes and reflector arrangement.(b) Current distribution when antenna works in VP1 state.

FIGURE 6 .
FIGURE 6. Patterns with and without choke plate and choke plate structure.

FIGURE 8 .
FIGURE 8. Comparison of isolation with and without the isolation structure.

FIGURE 9 .
FIGURE 9. Two views of the WLAN antenna prototype.

FIGURE 10 .
FIGURE 10.The isolation between HP antenna and VP antenna.

FIGURE 13 .
FIGURE 13.The radiation pattern of HP1.(a) Simulated radiation pattern.(b) The cutting plane corresponding to the maximum gain.

FIGURE 14 .
FIGURE 14.The radiation pattern of VP2.(a) Simulated radiation pattern.(b) The cutting plane corresponding to the maximum gain.
) and Fig.14(a) shows the 3D radiation patterns of the two polarizations.It can be seen from Fig.13that the angle of maximum gain of HP1 state is about ϕ = 20 • and θ = 65 • .Therefore, in order to obtain the maximum gain in the measurement, the radiation pattern of HP1 state is measured in the cutting plane of ϕ = 20 • , as shown in Fig.13 (b).Similarly, in Fig.14 (a), the angle of the maximum gain of VP antenna is about ϕ = 45 • and θ = 70 • , and in order to obtain the maximum gain, the radiation pattern of VP2 state is measured in the plane of ϕ = 45 • , as shown in Fig.14 (b).Thus it is indicated that all the beams of HP and VP antennas are slightly uplifted due to the effect of the metal ground on substrate5 with