Co-Circularly Polarized Planar Antenna With Highly Decoupled Ports for S-Band Full Duplex Applications

This work presents a unidirectional, co-circularly polarized (CP), printed antenna with highly decoupled or isolated transmit (Tx) and receive (Rx) ports for 2.4 GHz in-band full duplex (IBFD) applications. The presented antenna topology is based on four similar and sequentially rotated trimmed patches with right hand circular polarized (RHCP) characteristics. The symmetrical placement of two Tx patches with respect to both Rx elements results in equal levels of self interference (SI) which was suppressed through balanced excitation of Tx mode. This mechanism results in effective suppression of SI at each Rx patch. The residual SI is suppressed further through a second balanced feeding network deployed at Rxport of proposed antenna topology. The employed balanced feeding networks provide superior performance of ≤ 0.5 dB and 6° magnitude and phase imbalances respectively between the two balanced output ports over the bandwidth of interest. The measured results for prototype of presented antenna achieve -10 dB bandwidth of better than 100 MHz for both Tx and Rxports. The measured interport coupling for validation model ≤ −70 dB across the entire bandwidth of 100 MHz. As per best of authors’ knowledge, the presented antenna is the first one to report such reduced levels of interport coupling over the whole impedance bandwidth of planar antenna with unidirectional radiation patterns and co-RHCP characteristics for both Tx and Rx modes across the overlapped bandwidth.


I. INTRODUCTION
The future communication systems require novel and efficient duplexing techniques to offer higher data rates through improved spectral efficiency for wireless links [1], [2]. Such duplexing schemes should utilize the available bandwidth effectively to achieve higher throughputs [1], [2]. The in-band full duplex (IBFD) scheme also known as single frequency full duplex (SFFD) can theoretically double the spectral The associate editor coordinating the review of this manuscript and approving it for publication was Hassan Tariq Chattha . efficiency or link capacity through concurrent transmission and reception over the same bandwidth [3]- [6]. However, such simultaneous transmission and reception operation across the overlapping bandwidth results in strong co-channel coupling or self interference (SI) between the transmit (T x ) node and co-located receive (R x ) node. Due to such strong inband coupling, the SI power levels are much higher than those of the desired signal levels [3], [4]. Such strong co-channel interference degrades the capacity of R x channel through reduced signal to interference ratio for IBFD links [6]. In fact, the strong in-band SI signals act as jamming signals to VOLUME 10, 2022 This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/ overpower the desired R x signals [7]. The resulting in-band coupling or SI is comprised of both direct T x power leakage to R x channel and coupling of reflected T x signals (from surrounding objects) to its own R x node [8]. Furthermore, the former type of complex SI is comprised of both linear and non-linear components of T x signals [8].
The successful realization of IBFD operation requires the effective suppression or mitigation of SI at R x node to provide the intended inter-node isolation levels [7], [8]. Due to presence of very strong T x signals, significant levels of self interference cancellation (SIC) is required on R x node to provide the intended isolation levels for effective retrieval of low powered R x signals [7], [8]. The ideal case will be the suppression of SI signals to the inherent noise floor of receiver so that the desired signals from remote transmitter can be detected successfully [7], [8]. Practically, the intended SIC levels are defined by the T x power, noise figure and desired bandwidth of the IBFD transceiver [7]. The required high levels of T x -R x isolation demand the SIC operation at multiple stages (including frontend of transceiver and digital base band stage) across the IBFD transceiver [8]- [10]. Moreover, it is essential to achieve higher SIC levels at transceiver's front end (antenna stage and RF domain) comparative to other SI suppression stages in order to preserve the dynamic range of ADC for desired R x signals [11], [12]. This will also alleviate the required isolation levels imposed on later SIC stages of transceiver.
So far, various SIC techniques have been investigated to clinch elevated levels of interport isolation for antennas intended for IBFD applications [13]- [25]. The SIC techniques include the path loss decoupling based on spatial separation of antennas elements [13], isolation through uncorrelated dual polarized T x and R x signals [14]- [19], suppression of SI through near field cancellation i.e. near field SIC [20]- [22], and other topologies based on characteristics modes and different polarizations for T x and R x modes [23]- [25]. The path loss based isolation termed as spatial duplexing acquires the decoupling between T x and R x ports of IBFD antennaby increased spacing between T x and R x antenna elements or through separation of very narrow directional beams from respective antennas.
The dual polarization based isolation techniques exploit the intrinsic decoupling of orthogonal T x and R x signals to achieve up to 40 dB and 20 dB port to port isolation for linearly-polarized [18]- [20] and circularly-polarized antennas [15]- [17] respectively. Moreover, the dual polarized stacked printed antennas based on hybrid feeding networks can offer around 20 dB additional levels of isolation on the top of polarization isolation [26]. The interport isolation levels for such dual polarized antennas can be improved further through analog or RF domain SIC stages based on signal inversion techniques [26]- [28]. However, the achievable SIC bandwidth through such techniques is limited to few MHz and can be improved on the cost of additional complexity and power loss [27]. Furthermore, the polarization based decoupling techniques restrict the IBFD transceivers to employ the polarization duplexing for bidirectional communication where the forward and reverse links use different polarizations. Moreover, the polarization based duplexing is not viable in wireless systems intended for several applications including the continuous-wave radar systems based on shared or single antenna architectures [29]- [31]. In addition, the bidirectional wireless systems employing the same polarization for forward and reverse channels to safeguard the channel reciprocity to facilitate the second polarization for an additional full duplex link [32]. For instance, the linear vertical polarization can be used for first bidirectional link while the second full duplex link can utilize the linear horizontal polarization [33]. The intrinsic isolation of polarization duplexing will provide sufficient levels of isolation between two bidirectional links in that case. However, the polarization mismatch loss may still arise due to misalignment of co-polarized antenna employed at respective T x and R x nodes. The resulting polarization mismatch loss can be avoided effectively through the cocircularly polarized (same CP polarization) antennas. In that case, the left handed circular polarization (LHCP) or right hand circular polarization (RHCP) can be used for each full duplex wireless link. However, the dual port, compact copolarized CP antennas with significant levels of interport isolation are essential for such links.
Recently, the interport isolation techniques based on near field SIC topologies have been emerged to achieve excellent isolation levels through compact printed antennas [33]- [37]. Such techniques are equally effective for single antenna element [33]- [36] and antenna array configurations [37]. Moreover, these techniques provide nice SIC performance both for shared antenna and separate antenna architectures. Furthermore, the near field SIC techniques are well efficient both for dual polarized and co-polarized antennas to achieve either directional or omnidirectional radiations. However, the achievable isolation performance of near field SIC techniques rely on the characteristics of the employed feeding networks and symmetry of R x elements/ports with respect to T x element/port [33]- [37].
This work presents an antenna array configuration based on four trimmed patches to achieve right handed circular polarization and unidirectional radiation patterns for both T x and R x modes. The well-balanced feeding networks for each pair of T x and R x antenna elements achieve very high levels of port to port isolation through near field SIC operation. The employed balanced feeding networks achieve the high interport isolation without degradation in radiation features of presented co-circularly polarized antenna array. The structure of each feeding network is comprised of two inverted microstrip arms as output balanced ports and input unbalanced port excited through H-shaped slot in ground plane. The proposed IBFD antenna array and feeding networks have been implemented and interconnected to record the measurement results.
The rest of this paper is organized in following way: The section II presents the topology of proposed antenna array based on two pairs of T x and R x patches. The passive SIC operation for the case of ideal balanced feeding is described mathematically in this section too. This section also provides the full-wave simulation results for the proposed IBFD antenna with co-RHCP characteristics. The implementation details for the validation model of proposed antenna array are presented in section III. The novelty and contributions of this work are detailed in section IV. Followed by, the conclusions in section V.

II. ARCHITECTURE OF PROPOSED Co-CP ANTENNA ARRAY WITH PASSIVE SIC MECHANISM
The geometry of the proposed co-RHCP antenna array and its dimensions are presented in Fig. 1 for the operating frequency of 2.4 GHz. The architecture of presented antenna is based on four identical and sequentially rotated patches with RHCP characteristics. As depicted in Fig. 1, the four sequentially rotated antenna elements or patches have been placed symmetrically with respect to each other. Each patch is corner-trimmed to generate RHCP through single feed or port. As indicated in Fig. 1, two diametrically opposite pairs of RHCP patches have been used for differentially excited (balanced-fed) T x and R x . The balanced feeding operation will be realized through a compact and wideband slotcoupled balanced feeding network. As indicated in Fig. 1, a single layered FR-4 substrate with 1.6 mm thickness, relative permittivity (ε r ) of 4.4 and loss-tangent (tanδ) = 0.02; has been used for design and implementation of presented co-CP antenna array.
The trimmed-corner, square-shaped T x1 patch radiates a RHCP wave when excited from port 1 + through a thin coplanar feed line. The radiated RHCP wave is comprised of two orthogonal and equi-magnitude components of E fields. The phasor form representation of E fields from both pairs of T x and R x patches can be expressed as follows: where x ∧ and y ∧ represent the unit vectors along x and y dimensions respectively and E o denotes the peak amplitude of radiated electric field (E). For balanced-fed or differentially excited pair of T x ports (1 + and 1 − ) of antenna presented in Fig. 1, the resultant electric field vector (E T ) can be computed as: Similarly, the resulting electric field vector (E R ) for differentially excited R x mode through pair of R x ports (2 + and 2 − ) of proposed antenna array is given as: As evident from (3), the balanced or differential feeding through pair of T x ports of proposed antenna results in constructive combination (in-phase addition) of electric fields generated from each T x patch. Similar is the case for the incident (received) electric fields through the differential output operation for pair of R x ports of antenna presented in Fig. 1 as clear from (4). Furthermore, the radiated (transmitted) and incident (received) signals from each pair of balanced-fed or differentially excited T x and R x patches are right hand circular polarized (RHCP) as evident from the resulting E fields expressed by (3) and (4). The co-RHCP characteristics of presented antenna are also validated by the full-wave simulations through Ansoft HFSS software. The simulated surface currents densities and three dimensional radiation patterns for differentially excited T x and R x modes are exhibited in Fig. 2 to endorse the in-phase addition of E fields for each pair of patches. VOLUME 10, 2022 The presented antenna architecture provides very low levels of coupling (high isolation) between the single-ended T x and R x ports due to effective cancellation of SI at designated R x port. The reduced interport coupling mechanism based on balanced feeding at T x and R x ports can be investigated through a very simple analysis. As evident from the geometry of proposed antenna in Fig. 1, the symmetrical placement of both R x elements with respect to T x elements generates same levels of SI (in-band coupling from T x to R x patches). Moreover, the two components of SI at each R x patch have phase difference of 180 • degree due to the differential excitation of T x patches. Assume that P 11 , P 12 P 21 and P 22 denote the SI levels (in dB) for each pair of T x -R x patches. The SI power (P SI ) at port 2 + and 2 − of both R x patches can be expressed as follows: Based on the symmetry of the presented antenna structure P 11 = P 12 = P 21 = P 22 and SI should be completely suppressed (infinite isolation) at the output of each R x patch in ideal case. However, in practice P 11 ∼ = P 12 ∼ = P 21 ∼ = P 22 to offer finite levels of isolation between each pair of T x -R x patches and achieved isolation levels are dependent on balancing characteristics of employed feeding network for differential excitation of T x mode. Furthermore, the isolation levels can be improved by suppressing the residual SI (P SI 2+ and P SI 2− ) through differential R x operation. This mechanism will improve the isolation levels through following SIC operation where the SI power (P SI ) at singleended or unbalanced port of differential circuit can be expressed through following relation: As P SI 2+ and P SI 2− powers are already very low so the second stage of suppression will provide very high levels of isolation between single ended or unbalanced ports T x and R x ports of presented co-RHCP antenna structure.
It is evident from above analysis that the differential excitation for each pair of T x and R x patches can offer very high interport isolation levels without compromising the radiation performance of the proposed antenna. However, the achievable isolation levels for the presented antenna structure are highly dependent on the in-band amplitude and anti-phase balancing characteristics of the differential feeding networks. Moreover, the well-balanced T x and R x modes can also provide the improved gain performance for presented antenna through low side lobe levels (SLL).
For instance, the ideal balanced feeding networks can achieve ≥80 dB isolation across the entire −10 dB bandwidths of T x and R x ports as clear from the simulation results presented in Fig. 3. As evident from these simulation results, the −10 dB bandwidth for both T x and R x ports is better than 100 MHz and free space path loss based isolation between each pair of T x -R x patches is around 27 dB across the 100 MHz bandwidth. Based on (5), an ideal differentially excited T x mode improves the isolation levels to ≥80 dB across the bandwidth of interest. The interport isolation levels of better than 100 dB are achieved through an ideal differential-fed R x mode which cancels the residual SI as evident from (6). However, the resulting isolation levels through practical balanced feeding networks will be degraded due to non-ideal response of such networks.
It is evident from above discussions that the performance (amplitude and phase balances) of the balanced feeding network defines the achievable SIC levels for presented antenna structure. In this work, a modified version of a differential circuit (balun) reported in [38] has been used for balanced excitation of presented antenna. Two such symmetrical baluns have been used for balanced excitation of T x and R x modes. The geometry of modified wideband balun and its dimension are given in Fig. 4. The presented balun realizes the desired differential characteristics through a pair of microstrip to slot-line transitions. The narrow H-shaped slot-line in the ground plane provides the required coupling between the unbalanced port and the pair of balanced ports. The presented balun circuit is designed and implemented on a 1.6 mm thick FR-4 substrate with permittivity (ε r ) of 4.4 and loss-tangent (tan δ) = 0.02.
The geometrical dimensions of microstrip to slot-line transition based balun (balanced network) are detailed in Fig. 4. The simulation results like reflection coefficients, magnitude and the phase balance characteristics of proposed balun are detailed in Fig. 5. The port 1 is the unbalanced port while port 2 and port 3 are balanced ports of proposed balun. As clear from these simulated characteristics, the presented balun or balanced feeding network exhibits very low levels of reflections (well matched ports) for unbalanced and pair of balanced ports and -10 dB bandwidths are in excess of 1 GHz. As clear from   to 2.50 GHz. The insertion loss of the presented balun is less than 0.5 dB across the intended bandwidth as clear from Fig. 5(a). These improved characteristics of balun offer the potential of achieving high SIC levels through balanced T x and R x excitations of presented co-RHCP antenna.

III. EXPERIMENTAL CHARACTERISTICS OF PRESENTED Co-RHCP ANTENNA SYSTEM
In order to ensure the T x , R x ports matching, interport isolation (decoupling) performance and validate the far-field characteristics of the presented antenna system, a prototype (validation model) of antenna system was implemented using basic printed circuit board (PCB) technology. The antenna array and two baluns were implemented on FR-4 substrate having ε r = 4.4, tan δ = .02 and thickness of 1.6mm. The validation models of antenna-elements and both of the baluns is shown in Fig. 6. The sub-miniature (SMA) connectors have been mounted / soldered on designated ports of antennaelements (on back side of patches) and both baluns. The T x and R x baluns will be connected to respective ports of antenna-elements through phase-matched cables as indicated in Fig. 7(a).
The validation model was characterized through measurements with a calibrated vector network analyzer (VNA) shown in Fig. 7(a) to record its T x and R x ports matching and interport isolation parameters The simulated and measured S-parameters for presented antenna system are depicted in Fig. 7(b). Note that the simulated S-parameters for the antenna system were obtained through the interconnections of EM models of antenna-elements and baluns in schematic. This implies that the effects of phase-matched cables and SMA connectors were not considered in simulations. As indicated in Fig. 7(b), the presented antenna system offers −10 dB bandwidth in excess of 100 MHz which spans over 2.40 GHz to 2.50 GHz for both T x and R x ports. In addition, the assembled antenna system offers better than VOLUME 10, 2022 70 dB isolation levels over the entire matching bandwidth of 100 MHz as clearly reflected from measured results. As demonstrated earlier in simulation results, the path-loss isolation through spatial separation of each pair of T x -R x patches is around 27 dB and balanced feedings at both T x and R x ports elevate these isolation levels to 70 dB over the 100 MHz bandwidth. That means the balanced feeding contributes more than 43 dB isolation levels on the top of the spatial domain isolation.
As clear from Fig. 7(b), the simulated and measured Sparametrs closely follow each other except some significant deviations for the interport coupling results. These common deviations are based on the fact that the effects of connectorss and cables were not considered in simulations. Moreover, the fabrication accuracy and measurement tolerances also contribute to these deviations. However, the significant difference in simulated and measured interport coupling results is maninly due to additional magnitude and phase imbalances induced by cables used for interconnections of  antenna elements and both baluns. Another factor which results in degradation of measured interport coupling levels is due to reflection from surrounding metallic objects in lab environment. Finally, the minor frequency shifting for measured S-parameter results is attributed the tolerances of the dielectric parameters like thickness and permittivity etc.
The far field characteristics of validation model of presented full duplex antenna system were endorsed through gain measurements at operational frequency of 2.45 GHz. These gain levels for both modes have been measured through two antenna method where the antenna under test had been used as a receiving antenna. The measured gain levels for = 0 • , 45 • and 90 • far-field cuts are depicted in Fig. 8 and Fig. 9 for T x and R x modes respectively. The gain measurements performed for = 0 • , 45 • and 90 • (azimuth plane) are intended to endorse the CP characteristics of the prototype. Due to unidirectional radiation characteristics of the presented antenna, the gain measurements for other hemisphere (θ = 90 • to 180 • ) are not required. To record the gain levels for each port, the other port of the antenna system is terminated in 50 matched load to avoid the reflections from that port. The measurement results depicted in Fig. 8 for T x mode of antenna prototype offer better than 6.9 dBc peak gain at boresight (θ = 90 • ) for = 0 • , 45 • and 90 • cuts. Similarly, the recorded peak gains are also better than 6.8 dBc for = 0 • , 45 • and 90 • cuts when the R x port is used to record the power levels as clear from Fig. 9.
The balanced feeding approach offers better gain performance compared to single-ended excitation through reduced side lobes levels (SLL) i.e. reduced mutual coupling as reported in various earlier published works. The high loss FR-4 dielectric (tan δ = 0.02) can be replaced with low loss substrate to improve the radiation efficiencies and the resulting gains for both T x and R x modes. The resulting improvements in T x and R x gains can be validated through the simulations results for proposed antenna system by simply lowering the loss tangent values. For instance, 3 dB additional gain will be achieved if the radiation efficiency is doubled for antenna. It is important to note that both T x and R x modes have similar peak gain levels, same polarizations i.e. RHCP and overlapping -10 dB bandwidths along with high interport isolation levels as clear from measurements.
The radiation efficiency results for validation model (prototype) of presented antenna array were measured with the help of EMSCAN RFxpert near-field measurement setup. The measured radiation efficiencies are better than 53% for both T x and R x modes as clear from results presented in Fig. 10. The radiation efficiencies of the presented antenna can be improved through the use of low loss substrate for realization of prototype of this antenna.
The Fig. 11 depicts the simulated and measured axial ratio versus elevation angle (θ) results for presented antenna. The validation model of presented IBFD antenna characterizes 3dB axial ratio (|AR| ≤ 3dB) beam-width of 60 • for both T x and R x ports as clearly demonstrated through measured  results in Fig. 11. Moreover, the recorded values of minimum AR are 2.6 dB and 2.8 dB for T x and R x modes respectively. These experimental characteristics for implemented antenna demonstrate nice CP characteristics over wider beam-width in elevation plane for both modes.
The simulated and measured AR versus frequency results for the presented antenna are given in Fig. 12. As obvious from these results, the validation model of presented IBFD antenna exhibits |AR| ≤ 3dB over the entire matching bandwidth of 100 MHz (2.40 GHz to 2.50 GHz). These AR versus frequency measurements endorse the intended CP characteristics over the entire matching bandwidth of antenna for both T x and R x ports.

IV. NOVELTY AND CONTRIBUTIONS OF PRESENTED ANTENNA SYSTEM
The performance comparison of the presented antenna system with some of the previously reported co-CP antennas VOLUME 10, 2022 is detailed in table 1. The selected antennas for comparison purpose also utilize near filed SIC to improve interport isolation levels. The validation model (prototype) for our presented co-RHCP antenna in this work features very high levels of port to port isolation levels compared to previously reported co-CP antennas intended for full duplex applications. These elevated interport isolation levels are ascribed to effective suppression of resulting SI and complex interelement coupling through well-balanced feeding networks employed at both T x and R x ports of presented antenna with same polarization for both modes.
Compared to earlier reported co-CP antennas, the novelty of antenna system presented in this work is high interport isolation across the entire -10 dB impedance bandwidth of antenna along with same gain levels for both T x and R x modes. These high isolation levels along with RHCP polarization characteristics for both T x and R x modes offer the real gains of full duplex wireless operation without polarization-duplexing for bidirectional links. Moreover, the CP based wireless links avoid the inherent polarization mismatch losses in wireless links with linearly polarized characteristics. Furthermore, the presented antenna array offers better gain performance compared to single-element based IBFD antennas which will improve the coverage or range of wireless communication for given T x power.

V. CONCLUSION
A co-RHCP printed antenna with unidirectional radiation patterns is presented which achieves very high interport isolation over entire impedance of antenna for S-band full duplex applications. The superior in-band amplitude and outof-phase balance performance of feeding networks features excellent passive SIC operation for presented antenna array with co-circularly polarized characteristics. The employed balanced feedings also result in reduced mutual coupling between patches to achieve better gain levels through effective lower side lobe levels (SLL). The performance of the presented antenna prototype has been endorsed through measured impedance bandwidths for both T x and R x ports, port to port coupling levels, axial ratio versus frequency, axial ratio versus elevation angle parameters, and resulting antenna gains for both modes. As the insertion loss of the employed balun is very low (≤0.5 dB) across the intended bandwidth so the antenna gain performance will not be degraded when the baluns are used for balanced excitations of both modes. The gain of the presented antenna can be enhanced through implementation of given structure on low loss substrate which will offer improved radiation efficiency.  IEEE TRANSACTIONS ON  VEHICULAR TECHNOLOGY, the IEEE TRANSACTIONS ON MICROWAVE THEORY AND  TECHNIQUES, the IEEE WIRELESS COMMUNICATIONS, the IEEE TRANSACTIONS ON  AEROSPACE AND ELECTRONIC SYSTEMS, the IEEE TRANSACTIONS ON INDUSTRIAL  ELECTRONICS, and IEEE ACCESS.