Search

Artificial intelligence (AI) provides opportunities to enable high-efficiency wireless communication to dynamically adapt to the local environments and user demands. In this article, a continual learning digital predistortion (DPD) algorithm is proposed to linearize radio frequency (RF) power amplifier (PA) in 6G AI-empowered wireless communication. It is achieved by employing the ability of knowledge retention and knowledge transfer to merge the new PA operating state into the previous model, which makes the model more intelligent when dealing with multiple states. In addition, an augmented self-sensing model is proposed to accurately capture the common characteristics and key dynamic characteristics among different PA operating states to improve the knowledge transfer ability during the continual learning process. The experiments were carried out on a Doherty PA with dynamic configurations associated with power, bandwidth, and signal type. The experimental results have shown that the proposed technique can continuously learn when faced with unpredicted new PA operating states and so as to linearize all known states. The proposed DPD method performs with very low complexity in the long term, which is very suitable for the emerging AI-empowered dynamic scenario. [ABSTRACT FROM AUTHOR]

Future wireless systems that take advantage of in-band full-duplex (IBFD) technology require that the consequential self-interference (SI) be sufficiently mitigated for successful operation. Many IBFD designs utilize a multitap RF cancellation stage that necessitates the optimal configuration of multiple weights to accurately replicate real-world multipath SI channels. The nonconvex error surfaces generated by these channels present different features that can be challenging to navigate for canceller tuning algorithms. This article introduces a novel adaptive learning rate approach that utilizes a target error and a gradient estimation that is progressively forgotten over time to accelerate convergence speed and maximize cancellation performance. The tuning characteristics of a prototype system using this algorithm to set the weights of a two-tap canceller were measured over a 20-MHz bandwidth centered at 2.5 GHz and produced an average of more than 35 dB of cancellation for various SI channels with tuning times that are significantly less than previous approaches. The performance of this unique method was also studied against other tuning techniques, and results indicate that it could supplant traditional adaptive learning approaches for future IBFD systems. [ABSTRACT FROM AUTHOR]

In this article, we present a new method to reduce the model adaptation complexity for digital predistortion (DPD) of radio frequency (RF) power amplifiers (PAs) under varying operating conditions, using pretrained transformation of model coefficients. Experimental studies show that the PA behavior variations can be effectively tracked using a small number of “transformed” coefficients, even with large deviations in its output characteristics. Based on this discovery, to avoid reextracting all the original coefficients every time when the operating condition changes, we propose to conduct a one-time off-line pretraining stage to extract the common features of PA behaviors under different operating conditions first. The online model adaptation process will then only need to identify a small number of transformed coefficients, which can result in a drastic reduction in the computational complexity of the model adaptation process. The proposed solution is validated by experimental results considering varying signal bandwidth and output power levels on a high-efficiency gallium–nitride Doherty PA, where the computational complexity is significantly reduced and the system performance is not compromised. [ABSTRACT FROM AUTHOR]

This article proposes a 28-GHz full-duplex (FD) phased arrays, which consists of 64-element transmitting and receiving arrays, and a 28-GHz radio frequency (RF) canceller. The phased arrays are designed with $2\times 2$ beamformer chips, and the transmitter (TX) and receiver (RX) arrays are oriented to have cross polarized configurations for high isolation. The 28-GHz RF canceller is based on an identical $2\times 2$ beamformer chip with two external different delay taps. Since the canceller is located at the input and output ports of arrays instead of antennas, it cannot eliminate the nonlinearities and noises in the phased arrays. Therefore, these nonlinearities and noise in the TX and RX arrays are experimentally analyzed to demonstrate that the proposed cross-polarized arrays have enough isolation to achieve suppression to the level of the RX noise floor. The FD phased-array system at 28.5–29.5 GHz has 57-dB isolation between the TX input and RX output ports, where the cross-polarization improves 10-dB isolation comparing to co-polarization and the RF canceller provide additional 10-dB self-interference (SI) suppression. Also, the nonlinearities and noise components of the received SI signals are lower than the noise floor at the output of the RX array. The system requires only 15.5-dB of additional digital cancellation to reduce the SI to the RX noise floor of 1-GHz bandwidth while maintaining the transmitting EIRP at 41 dBm. [ABSTRACT FROM AUTHOR]

This article presents a new circuit topology for a highly reconfigurable $1\times 4$ radio frequency network that, for a given input, is able to modify magnitude and phase relations among its output signals with great flexibility while preserving matching at all ports. The circuit, based on a four-way power divider, four hybrid couplers, and eight variable phase shifters, is first theoretically analyzed. It is shown that magnitude and phase relations among signals at the output ports can be modified just by changing the state of the variable phase shifters, while the sum of the squared magnitude of the transmission coefficients of the network is always equal to one. For the specific purpose of the experimental validation of the developed theory, a particular version of the network, working at 3.6 GHz, is then designed and tested. All configurations are affected by the same insertion loss, depending on the adopted phase-shifting technology. The obtained results show the potential of the conceived circuit and testify its unprecedented degree of reconfigurability. [ABSTRACT FROM AUTHOR]

This article presents a technique to suppress noise in switched-capacitor (SC) radio frequency (RF) receivers (RXs). In the proposed technique, the sampling capacitor of a typical SC-RX is realized using a MOS capacitor. Furthermore, the same sampling MOS capacitor is switched between the inversion and depletion regions to act as a discrete-time parametric amplifier for downconverted signals. Due to the voltage gain at the front end, the noise of successive stages in the receiver chain is suppressed. Analytical expressions for the conversion gain, input impedance, and noise figure (NF) of the gain-boosted SC receiver are derived and verified through simulations. A prototype SC receiver is implemented in a CMOS 180-nm technology. In measurements, the receiver achieved 41-dB conversion gain and 3.1-dB NF over the tuning range of 0.2-to-0.95 GHz. The reduction in NF due to the parametric gain boost is >10 dB. [ABSTRACT FROM AUTHOR]

This article presents a frequency reconfigurable RF impedance tuner using eight cascaded fluidic cells to cover the frequency band [0.9, 2.4] GHz. The tuner is made of eight fluidic fillable cavities on top of a coplanar waveguide transmission line in a single low-temperature cofired ceramic (LTCC) substrate. Each cavity is either kept empty or filled with deionized (DI) water to alter the substrate’s permittivity and so change the tuner’s response. Closed-form expressions are derived and used to dimension the tuner’s individual cells, while 3-D field simulations are used to complete the design of a single cell and an entire eight-cell tuner. Both components are fabricated, and good agreement between simulation and measured is observed. The fabricated tuner is used to build a fluidic reconfigurable RF amplifier. [ABSTRACT FROM AUTHOR]

In this article, we present a dynamic model sizing approach for digital predistortion (DPD) of broadband radio-frequency power amplifiers. By employing a novel model structure adaptation algorithm, the DPD model structure can be adaptively adjusted during its real-time deployment to keep the optimum size and complexity under different operation conditions. Power consumption of DPD can be reduced by on-demand automatic model structure adaptation instead of reusing the same model structure for all power levels and band allocations. To realize dynamic model sizing, the adaptation algorithm explores new potential terms based on prior knowledge of the model structure and prunes the DPD model with a stepwise backward regression method. Experimental results show that the algorithm can quickly find the optimum model structure when the operation condition changes. During the adaptation, it can also maintain robust linearization performance with a relatively low computational complexity and thus demonstrates itself as a suitable solution to the linearization of future broadband wireless systems. [ABSTRACT FROM AUTHOR]

In this article, we study the propagation of high-power RF fields along an air channel realized in a dielectric photonic crystal (PC) woodpile. The resulting hollow-core dielectric waveguide is suitable for high-power applications at 18 GHz and can be used as a dc-break for microwave ion sources. The infinite-periodic electromagnetic bandgap (EBG) waveguide structure has been truncated and coupled to a standard metallic rectangular waveguide operating in the fundamental TE10 mode. The fabricated waveguide structure exhibits a very low insertion loss (IL) of 0.5 dB over a transmission bandwidth of 260 MHz around the operating frequency of 18.25 GHz. A six-period structure has been manufactured, and the experimental characterization exhibits a very good agreement with the simulation results. [ABSTRACT FROM AUTHOR]

A new, frequency-domain, behavioral modeling methodology for radio frequency (RF) power transistors, based on canonical section-wise piecewise linear (CSWPL) functions, is presented in this article. The basic theory of the proposed model for fundamental, harmonic, and dc responses is provided within. Compared with the existing nonlinear behavioral models, which require tables of coefficients to account for varying input powers, the model described in this article is able to predict transistor behavior over the entire Smith chart, at different levels of input power, from the linear region to strongly nonlinear region, with only a single set of model coefficients. Model verification is performed at fundamental, harmonic, and dc frequencies through comparisons with simulated data from a reference nonlinear circuit model and with experimental data from separate 10- and 2-W GaN HEMT devices, over a wide range of load conditions. The models can predict, accurately, the optimal area on the Smith chart for maximum fundamental output power. Furthermore, the ability of the model to interpolate across input power levels is also tested, with excellent fidelity to the simulated and measured data obtained. Compared with a 2-D polynomial-based model, the proposed method provides for a more accurate prediction, while using fewer model coefficients. [ABSTRACT FROM AUTHOR]

This article presents a quadrature balanced radio-frequency (RF) front-end transmitter (TX) architecture assisted by a digital self-interference cancellation (SIC) algorithm for simultaneous transmission and reception in next-generation full-duplex radios. The proposed quadrature balanced power amplifier (QBPA) topology is a four-port circuit with an embedded passive primary isolation (PI) between the antenna and the receiver (RX) low-noise amplifier (LNA), along with a very low power SIC injection mechanism that reuses the TX’s gain. A transmit–receive (TR) isolation of >57 dB is measured for continuous-wave (CW) signals over an 80 MHz bandwidth (BW). A TR isolation of >50 and >56 dB is demonstrated for an 80 MHz 802.11ac orthogonal frequency-division multiplexing (OFDM) signal at 10- and 20 dB backoff from 30 dBm peak TX power ($P_{\text {TX,max}}$), respectively, within the 2.4 GHz Wi-Fi band. A digital predistortion algorithm is applied to linearize the TX at the high output power case, achieving error vector magnitude (EVM) lower than −34.7 dB for concurrent TX-SIC operation. A simultaneous TR scenario is demonstrated, measuring an RX signal EVM of −22 dB, while $P_{\text {TX,max}}\,\,=$ 30 dBm. Measured RX loss is lower than 1.8 dB within the operating frequency band. [ABSTRACT FROM AUTHOR]

This article reports signal processing and circuit techniques for direct RF signal modulation. A pulse-encoded transition (PET) technique is introduced to reduce undesired harmonic distortion (HD) and adjacent channel leakage ratio (ACLR) generated by the direct RF modulation. To enable PET RF switching, two variations of a high-power, stacked-FET switch modulator, an 8- and 12-device stack, are designed in 45-nm silicon on insulator (SOI) CMOS. The switch design uses a tapering design to significantly improve power handling with minimal impact to switching speed. The modulators have $P_{1\,{\mathrm{ dB}}} $ values between 34 and 39 dBm while demonstrating a modulation bandwidth of nearly 500 MHz with a 1-GHz carrier. The input referred third order intercept point (IIP3) is between 46 and 61 dBm. Additionally, ACLR measurements of up to −50 dBc are demonstrated using the proposed PET technique at 30-dBm output power. To the best of our knowledge, this is record power handling and ACLR for a CMOS switch. [ABSTRACT FROM AUTHOR]

This paper explores different functionalities that can be implemented directly over the optical layer, paving the way for the future for converged wireless and wired services optical networks. Wireless technologies operating over optical networks mainly employ radio-over-fiber (RoF) technologies, which effectively help to manipulate microwave and millimeter-wave band signals. We have experimented with different generation technologies in order to achieve distribution of microwave and millimeter-wave signals operating at different bands, ranging from a few gigahertz up to 60 GHz. We will show how these technologies, in combination with novel devices, can be used to enable networks in which dynamic channel allocation and peer-to-peer are carried entirely in the optical domain. Furthermore, since wireless and wired services fall within the domain of optical-access networks, we will demonstrate the performance of basic functionalities, such as distribution, transmission, and recovery at the local-access points. [ABSTRACT FROM PUBLISHER]