By Topic

IEEE Quick Preview
  • Abstract



The tremendous demand for broadband wireless communication applications, combined with the limited available radio frequency (RF) spectrum, has driven the global demand for increased capacity and data rates of wireless networks to accommodate larger volumes of subscribers. However, in current wireless network systems, the base station has no information on the position of each mobile user, and radiates the RF signal in all directions within a cell to provide radio coverage. This results in inefficient utilization of the radiated power during the transmission and causes interference to adjacent cells (cochannel) that use the same frequency. In addition, the antenna receiver detects signals from all directions including noise and interference signals, making the processing of the desired signals complicated, thus limiting the transmission speed and the number of users [1].

A phased-array antenna is an attractive solution that can be used to improve spectrum usage and increase network capacity, together with a more efficient use of transmitted energy. To obtain an optimal radiation pattern for broadband transmission, the signals received by or transmitted from the antenna array must be accurately time-compensated via a true-time RF delay generation mechanism [2], [3], [4], [5]. Electronic phase-shifting devices that are currently used for beamforming are frequency dependent, whereas conventional RF transmission delay lines are limited by the loss of their metallic media, especially at high frequencies. On the other hand, a smart antenna system based on the use of digital signal processing (DSP) is currently thwarted either by the limited resolution and the narrow bandwidth of analog-to-digital converter, or by the high power consumption of broadband analog-to-digital converters. Compared to all-electronics beamforming, optical RF signal processing offers attractive features such as small size, low weight, immunity to electromagnetic interference (EMI), compatibility with optical fiber network and, especially, the capability to generate true-time delays (TTDs) thus realizing squint-free broadband beamforming [6], [7], [8].

Tunable TTD generation can be made using free-space optics [9], integrated optics [10], or fiber optics [11], [12], [13]. Most TTD technologies are based on either path selection or fiber dispersion. The path selection approach generally provides discrete time delays, while the dispersion-based approach has advantages of generating continuously tunable time delays, thus enabling both beam and null steering.

Several approaches have been adopted to realize dispersion-based tunable TTD units for photonic beamformers, including the use of Bragg gratings (FBGs) [14], high dispersion fibers [15], dispersion-enhanced photonic-crystal fibers [3], and higher-order mode dispersive multimode fibers [16]. However, none of the reported photonics-based TTD units has the flexibility to either tune the TTD continuously or generate multiple tunable TTDs for each antenna element simultaneously; this is crucial for null steering. Since current photonic beamformers are limited in flexibility, accuracy, and reconfigurability, the need for new photonic technologies that can realize practical, flexible and cost-effective broadband beam and null-steering beamformers has become a major research focus worldwide. Opto-VLSI is a promising emerging technology that can deliver practical broadband phased-array antenna beamformers and meet the requirements of future wireless systems [17], [18]. The Opto-VLSI technology is based on the integration of Microelectronics and Photonics, and can be used to overcome the flexibility and cost-effectiveness bottlenecks of traditional photonic beamformers. The combination of a software-driven Opto-VLSI processor and photonic components has been widely applied in adaptive high-speed optical communications, demonstrating large potential to attain broadband microwave beamforming operation with high-level of flexibility and reliability.

The authors have reported a TTD architecture for null steering based on an Opto-VLSI processor [6], where the capability to generate multiple dynamic TTDs for one antenna element was demonstrated. However, the generation of multiple dynamic TTDs for all the antenna elements is still challenging, since the spectral slice from the amplified spontaneous emission (ASE) source is usually wider than the wavelength spacing required by null steering for an antenna element.

In this paper, we adopt our recent progress in Opto-VLSI-based fiber lasers for realizing a TTD architecture, with the difference that single wavelength narrow linewidth lasing is used instead of spectrally slicing the ASE wavelength span [17]. We demonstrate, for the first time, a novel broadband beamsteering beamformer employing an Opto-VLSI processor, erbium-doped fiber amplifiers (EDFAs), and high dispersion fibers to simultaneously generate arbitrary TTDs for each antenna element. This beamformer open the way for the realization of adaptive (or smart) phased-array antennas that significantly increase the capacity of next-generation wireless systems. The proposed beamformer has a number of novel features. First, it can adaptively achieve broadband beamsteering, and has the potential to achieve null steering (with multiple wavelength lasing being generated for each antenna element) through software; second, it incorporates microelectronics and photonics (Opto-VLSI) for RF signal processing, thus adding the flexibility, tunability, accuracy, and reconfigurability of microelectronics to the broadband capability of photonics; and third, it provides a cost-effective and compressed-hardware solution to antenna beamforming in next-generation wireless systems.



An Opto-VLSI processor is an array of reflective nematic liquid crystal (LC) pixels controlled by Very-Large-Scale-Integrated (VLSI) electronic circuits on the backplane of the LC. Each pixel can be individually addressed by applying a discrete voltage level between a transparent indium-tin oxide (ITO) layer on top of the LC and the aluminum mirror electrode on the silicon substrate. A quarter-wave-plate (QWP) layer is deposited between the LC and the aluminum mirror to accomplish polarization-insensitive operation. By driving the Opto-VLSI processor with blazed-grating-like phase holograms, a light beam incident onto its active window can be steered by the angle Formula$\alpha_{\rm m}$ [17] Formula TeX Source $$\alpha_{m} = \arcsin\left({m\lambda \over d}\right) \eqno{\hbox{(1)}}$$ where Formula$m$ is the diffraction order, Formula$\lambda$ is the light wavelength in vacuum, and Formula$d$ is the grating period. In this paper, the optical beam steering is employed to simultaneously generate variable TTD for RF signals associated with four phased-array antenna elements.

When an RF signal strikes a phased-array antenna, the incremental time delay Formula$\tau$ induced between the adjacent elements depends on their spacing, Formula$d$, and the incidence angle, Formula$\theta$, that is Formula TeX Source $$c \cdot \tau = d \cdot \sin(\theta)\eqno{\hbox{(2)}}$$ where Formula$c$ is the speed of light. Through introducing a time delay Formula$\tau_{m}^{\prime}$ for each element, the phase difference between antenna elements can be changed, leading to a controllable combined RF signal, Formula$S_{sum}(t)$, expressed as Formula TeX Source $$S_{sum}(t) = \sum_{m = 0}^{N - 1}S_{m} \left(t - m\tau - \tau_{m}^{\prime}\right) \cdot \cos\left[\omega_{c}t - \omega_{c}\tau_{m}^{\prime} - m\omega_{c}\tau + \varphi\left(t - m\tau - \tau_{m}^{\prime}\right)\right] \eqno\hbox{(3)}$$ where Formula$S_{m}(t - m\tau)$ and Formula$\varphi(t - m\tau)$ are the amplitude and phase of the incoming signal received by the Formula$m$th element, Formula$m = 1, 2, \ldots, N - 1$, Formula$\omega_{c}$ is the carrier frequency, and Formula$N$ is the number of antenna elements. Equation (3) shows that the received signal could be arbitrarily synthesized by adjusting the time-delay, Formula$\tau_{\rm m}^{\prime}$, experienced by the signals received by the antenna elements, so that the antenna gain, known as spatial directivity or radiation patterns, along any direction can be dynamically changed.

Fig. 1 shows an experimental setup illustrating the principle of the proposed Opto-VLSI-based RF phased-array antenna architecture. The optical amplifiers Formula$(\hbox{OA}_{\rm i})$ were used as gain media to realize four tunable fiber lasers whose output wavelengths are controlled by a single 512 × 512-pixel Opto-VLSI processor of 256 phase levels and a pixel size of 15 Formula$\mu\hbox{m}$. The optical amplifiers were EDFA operating in C-band. The broadband ASE noise resulting from each optical amplifier was split by the optical coupler with a 5/95 power splitting ratio, where 5% of ASE power was used to extract the output of the tunable fiber laser while the remaining 95% was recirculated in the fiber ring cavity to generate lasing. The polarization controllers were used to optimize the diffraction efficiency of the grating plate and to enforce single-polarization lasing for all tunable lasers. All broadband ASE signals of the EDFAs were directed to the corresponding collimator array ports, via optical circulators, and collimated at about 0.5 mm diameter. An optical lens (Lens 1) of focal length 10 cm was used between the collimator array and a diffraction grating plate to focus the collimated ASE beams onto small spots onto the grating plate. The grating plate, which has 1200 lines/mm and a blazed angle of 70° at 1530 nm, spatially demultiplexed each ASE spectrum along different directions. Another optical lens (Lens 2) with the same focal length as Lens 1, located at the middle position between the grating plate and the Opto-VLSI processor, was used to collimate the dispersed optical beams and map them onto the surface of the 2-D Opto-VLSI processor, which was logically partitioned into 4 rectangular pixel blocks. Each pixel block was assigned to an antenna element and used to efficiently couple back any part of the ASE spectrum illuminating this pixel block along the incident path into the corresponding collimator port. The Opto-VLSI processor arbitrarily selected one of the wavebands in each ASE spectrum that were mapped onto its surface block using the principle of beam steering described above.

Figure 1
Fig. 1. Opto-VLSI-based phased array antenna beamformer for broadband beam steering.

A Labview software was especially developed to generate and upload the optimized digital phase holograms onto the Opto-VLSI processor for simultaneously steering a selected waveband from each port and injecting it back into the corresponding collimator for lasing, through the recirculating fiber loop. This structure enabled the generation of multiple continuous wave (CW) laser signals, whose wavelengths were independently tuned over the gain spectrum of the EDFAs. Each wavelength channel was consequently modulated by the RF signal received by the associated antenna element. In the experiments, four different wavelengths were independently selected for lasing within the corresponding fiber loops by uploading appropriate phase holograms (blazed grating) onto the various pixel blocks of the Opto-VLSI processor. Each wavelength channel offered an output power of about 7 dBm, which was enough to compensate the overall insertion loss in the beamsteering beamformer. Note that the insertion loss caused by the passive optical coupler increases with the number of antenna elements. As an example, for a 100-element phased-array antenna, the insertion loss of the optical coupler is 20 dB, which can easily be compensated for by using a low-noise flat gain optical amplifier.

Each RF signal received by the element at the front-end of the phase-array antennas, with the amplitude and phase of Formula$S_{m}(t - m\tau)$ and Formula$\varphi(t - m\tau)$ as shown in (3), was used to intensity modulate the wavelength channels using electrooptic modulators (EOMs). All the RF-modulated optical signals were coupled into a single fiber and routed into a 10-km Corning LEAF nonzero dispersion shifted optical fiber with dispersion coefficient about 4.2 ps/nm/km and insertion loss of 0.2 dB/km at 1550 nm. Each RF-modulated optical signal experienced a TTD Formula$\tau_{m}^{\prime}$ that depended on its center wavelength, which was controlled by the Opto-VLSI processor. All RF-modulated optical signals were finally detected by a single photodiode that produced the sum of the delayed RF signals. This way, the TTD between adjacent antenna elements were generated by controlling the wavelength spacing between the various wavebands selected by the Opto-VLSI processor. One of the attractive features of the proposed phased-array antenna architecture shown in Fig. 1 is its ability to simultaneously generate and tune all the RF TTDs for the smart antenna beamformer through phase hologram generation. Various phase holograms were synthesized and optimized for specific beam steering scenarios, and stored to enable beamsteering scenarios to be recalled through software.

Note that the TTD resolution depends on the tuning resolution of the tunable laser sources as well as the dispersion of the delay-generating fiber. In the experiments, the 30-nm wide ASE spectrum of the optical gain medium was mapped onto the Opto-VLSI processor, which comprised 512 × 512 pixels. The wavelength tuning resolution, which mainly depends on the pixel size, was around 0.06 nm, resulting in a TTD resolution of 2.5 ps for the 10-km Corning LEAF nonzero dispersion shifted optical fiber used in the experiments. On the other hand, the maximum generated TTD is determined by the maximum tuning range of the laser sources, the number of antenna elements (N), and the dispersion coefficient of the delay-generating fiber. For a laser tuning range of 30 nm and an N-element antenna, only Formula$({\rm N} - 1)$ TTD are required to be generated and the maximum wavelength difference between the adjacent antenna elements is Formula$30/({\rm N} - 1)$ for RF beam steering, corresponding to a maximum TTD of 420 ps for Formula${\rm N} = 4$. From (2), a steering angle of 60° requires a maximum time delay of 288 ps, which can easily be achieved with our beamformer.



Experiments were conducted using the setup illustrated in Fig. 1 to evaluate the beam steering performance of a 4-element rectangular patch type smart antenna system. The antenna elements were separated by a distance of 99 mm, close to half of the operating RF signal wavelength Formula$(f_{c} = 1.85\ \hbox{GHz})$ thus alleviating the effect of sidelobes. By recalling the phase holograms generated for various scenarios, the lasing wavelengths associated to the antenna elements were tuned in such a way that their separations generated various incremental TTDs corresponding to different RF beam steering.

The experimental setup is illustrated in Fig. 2. A Log-periodic dipole antenna was used as a transmitter, by which a wide-angle pattern was generated. A 4-element phased-array antenna operating in a receive mode was placed at around 1 meter from the transmitter thus ensuring far-field operation. The signals received by the four antenna elements were amplified to modulate the four CW laser wavebands generated by the different tunable lasers. The RF-modulated laser signals were combined and fed into a photodetector through the nonzero dispersion shifted optical fiber, and an RF power meter was used to measure the generated RF signal. Note that, through adjusting the phase hologram the output power level for each channel can be adjusted to optimize the antenna directivity. The phased-array antenna was mechanically rotated 360° using a stepper motor with rotation step of 1°, and the received RF signal for each step was measured and recorded. A computer algorithm was developed to automatically measure the received RF intensity at different rotation angles.

Figure 2
Fig. 2. Setup for demonstrating the concept of Opto-VLSI-based broadband beamforming.

Four beamsteering scenarios were shown to demonstrate the principle of Opto-VLSI-based beamforming, by only changing the phase hologram uploaded onto the Opto-VLSI processor. Each phase hologram selected a specific laser wavelength for each antenna element, thus synthesizing a TTD profile that corresponded to an RF beam steering scenario. Fig. 3(a)(d) show the measured antenna radiation patterns (right) for several scenarios, each corresponding to the synthesis of wavelength channels (left) of different spacing. For example, in Fig. 3(a), the Opto-VLSI processor generated a wavelength spacing of 1 nm, which corresponds to 42-ps delay between adjacent antenna elements. The measured beamsteering angle was about 7°, which is in excellent agreement with the theoretical prediction of 7.2°. When the spacing between the laser channels was increased from 1 nm to 4 nm, the measured main lobe was steered from 7 ° to around 30°, as illustrated in Fig. 3. Note that, the increase in the sidelobe level from 7° to 30 °, is caused by parasitic effects, as explained by Burla et al. [19]. These experimental results agree very well with the theoretical predictions, and demonstrate the capability of the Opto-VLSI processor to realize a phased-array antenna beamformer. Also, the measured radiation patterns shown in Fig. 3 are in very good agreement with those reported by Burla et al. [19], who generated variable TTDs by using optical ring resonator TTD units in conjunction with a complementary metal–oxide–semiconductor (CMOS) based thermooptical tuning mechanism.

Figure 3
Fig. 3. Opto-VLSI-based tunable laser generations (left), and the corresponding RF beam steering profile measured in the experiments (right).

The main advantage of our Opto-VLSI-based Beamformer is its ability to employ a single Opto-VLSI processor, which can be mass produced, to simultaneously and independently generate multiple TTDs for not only beam steering but also null steering. Compared to other approaches to TTD generation, the dimensions of our proposed beamformer are not significantly affected when the number of antenna elements is increased. Only the size of the active aperture of the Opto-VLSI processor (which is a few centimeters) needs to be increased.

The generated tunable laser wavelength had a linewidth of less than 0.01 nm, which did not display mode competition, and therefore no deterioration of TTD beamforming performance was observed in the experiments. Note, however, that the laser linewidth may degrade the performance of the photonic beamforming if the operation frequency is increased beyond 10 GHz [20], [21]. This can be overcome by generating single mode lasing, through introducing techniques such as saturated absorption [22]. Finally, it is important to notice that the Opto-VLSI processor is a free-space optical system that requires careful alignment of various optical components. However, with advances in free-space optomechanical design and packaging, a compact and reliable TDD unit can be developed, as evidenced from many commercial Opto-VLSI based devices, such as reconfigurable optical add/drop multiplexers [23].



We have proposed and demonstrated the principle of an Opto-VLSI-based RF phased-array antenna employing tunable TTD units for adaptively steering the receive pattern. A single Opto-VLSI processor has been able to simultaneously and dynamically generate four independent wavelength channels and synthesize TTDs through a dispersion shifted optical fiber. Experimental results have shown the capability of the beamformer to steer the RF beam by up to 30°. The proposed smart antenna has application in Space Division Multiple Access (SDMA) networks.


Corresponding author: F. Xiao (e-mail:

F. Xiao, B. Juswardy, and K. Alameh are with the The Centre for MicroPhotonic Systems, Edith Cowan University, Joondalup WA 6027 Australia.

S. Xiao and W. Hu are with the State Key Laboratory of Advanced Optical Communication Systems and Networks, Shanghai Jiao Tong University, Shanghai 200240 China.


No Data Available


No Photo Available

Feng Xiao

No Bio Available
No Photo Available

Budi Juswardy

No Bio Available
No Photo Available

Kamal Alameh

No Bio Available
No Photo Available

Shilin Xiao

No Bio Available
No Photo Available

Weisheng Hu

No Bio Available

Cited By

No Data Available





No Data Available
This paper appears in:
No Data Available
Issue Date:
No Data Available
On page(s):
No Data Available
INSPEC Accession Number:
Digital Object Identifier:
Date of Current Version:
No Data Available
Date of Original Publication:
No Data Available

Text Size