A probe-fed patch antenna with 180° ring hybrid is proposed for in-band full-duplex applications. The proposed antenna consists of two substrate layers with the ground plane in between. A patch is printed on one substrate opposite the ground, and a ring hybrid is printed on a second substrate, which is designed to separate the receiving and transmitting signals, producing high isolation between the two. Simulated and fabricated prototype results show isolation of −58 dB covering the Industrial, Scientific and Medical (ISM) band of 2.4–2.5 GHz with a centre frequency of 2.45 GHz.

Millimetre wave (mmWave) communications provide a favourable solution to the ever-increasing data rate demands because of their abundant bandwidth availability. The conventional wisdom about the mmWave communication channel is that it always exhibits line-of-sight (LOS) propagation with directive antennas. However, in mobile communications, mmWave frequencies suffer from frequent LOS blockages and exhibit multipath propagation, which motivates the study and analysis of the polarimetric properties of the mmWave channel. While the exploitation of polarisation is promising, the cross-polarisation associated with it can be detrimental to the overall performance. Therefore, in this Letter, the authors propose a hybrid precoder design, which mitigates the cross-polarisation by the joint design of a radio-frequency beamformer as well as the precoder and combiner used in the baseband. The authors demonstrate through simulations that the proposed design outperforms eigen-beamforming by 10 dB at a given rate, when considering polarisation.

A dual-band antenna with good stopband characteristics is proposed. Different from conventional antennas, two transverse slots are etched on a substrate integrated waveguide (SIW) cavity. Two half modes are excited to achieve a dual-band operation and two non-radiating modes, a TE₁₁₀ mode and an even TE₁₂₀ mode, are induced to generate radiation nulls. With the aid of the two radiation nulls, the stopband characteristics of the proposed design including non-radiating gain and band-edge selectivity are significantly improved. A fabricated prototype is measured. The measured centred frequency of two passbands is at 3.28 and 3.77 GHz, respectively. The maximum gain is about 5.2 dBi in the passbands while the minimum gain is only −18.9 dBi in the stopband. The measured efficiency is 86% in the lower passband and 75% in the upper passband but it sharply decreases to lower than 5% between 3.36–3.7 GHz.

A simple internal antenna that combines a planar inverted F antenna (PIFA) and a loop antenna to achieve wideband (BW ∼ 48%) operation in ultra-high frequency (UHF: 470–771 MHz) band for digital television (DTV) applications is presented. The design allows the installation of an antenna within a TV model, thereby addressing the limitations of space. The 49-inch DTV is used for the practical implementation experiment. The TV bezel (made of polycarbonate-acrylonitrile butadiene styrene (PC-ABS): ε_r = 3.5; tan δ = 0.005) is used as the substrate for the antenna design. The antenna is placed on the top edge of the TV model. With the combination of the PIFA mode and loop mode operations, the operating bandwidth of the proposed PIFA-loop antenna is significantly broadened. The measured results show that the proposed antenna has wide bandwidth of about 410 MHz (460–870 MHz) and suitable radiation performance with higher than 2.2 dBi of peak gain and 47% of efficiency within the operating bandwidth.

A high sensitivity Class-3 ultra high-frequency (UHF) radio-frequency identification (RFID) system based on harmonic backscattering to achieve a long read-range is presented. Harmonic backscattering is a promising solution to solve the carrier leakage issue of UHF RFID reader to improve the sensitivity. In the proposed RFID system based on harmonic backscattering, the RFID reader transmits data with the fundamental frequency carrier and receives data with its second harmonic carrier, and the RFID tag employs a rectifier to generate both DC power and the second harmonic carrier for backscattering without impedance mismatch. The first harmonic Class-3 UHF RFID system based on National Instruments software defined radio platform is proposed and implemented. The proposed semi-passive harmonic RFID tag has dBm sensitivity. Test results show that the harmonic RFID system has a dBm sensitivity with the harmonic RFID tag whose signal-to-blocker ratio is dB at 20 m communication distance.

An inverter-based operational transconductance amplifier (OTA) robust to process, voltage and temperature (PVT) variations is proposed in this Letter to enable the design of low-voltage power-efficient continuous-time sigma-delta modulators (CT-SDMs). The proposed OTA significantly reduces the unity gain-bandwidth variation across PVT corners while presenting a DC gain of 41.9 dB at nominal conditions when operated at 0.6 V. The effectiveness of the proposed circuit was verified in a 180 nm CMOS process by transistor-level simulations including PVT variations and mismatch. Also, a 0.6 V third-order CT-SDM is designed to verify the robustness against voltage and temperature variations.

A DC current injection circuit of a trifurcated autotransformer (TAT)-based 12-pulse diode bridge rectifier (DBR) is proposed which can help reduction of harmonics in the AC mains. The circuit at the DC side employs a voltage multiplier (VM) circuits namely voltage doubler, voltage tripler and voltage quadrupler and an interphase transformer which provides the required circulating current. This induced circulating current shapes the input AC line current near to a sine wave. The output of VM circuits is connected across the load and thereby reuses the harmonic energy absorbed by the system. Thus, improves the energy conversion efficiency of the proposed system. The TAT-based 12-pulse DBR has the implicit ability to prevent the zero sequence components thus expels the necessity of zero sequence blocking transformer. The proposed configuration is analysed, simulated in MATLAB Simulink and the simulation results are presented, which confirms the improvement in power quality parameters in the input AC line current. Further, the viability of the proposed configuration is verified by experimental results which confirm the suitability of the proposed configuration in AC–DC applications.

A high-precision time integrator is presented whose output is the accumulation of sequence of time inputs. A novel time adder with time-domain differential structure is proposed to avoid common-mode errors such as supply noise and leakage current. The performance under various process, voltage and temperature conditions is examined in detail. By using two time adders in parallel, a differential synchronous time integrator is proposed, which is easier to cascade with other time-mode arithmetic circuit. Implemented in 65 nm 1.2 V CMOS technology, the proposed time integrator achieves a gain of 27.97 dB with a 317 kHz 30 ps peak-to-peak sinusoidal input, and consumes with 50 MHz sampling rate. The experimental results confirm the integration operation is realised properly and accurately.

A bipolar junction transistor (BJT)-based CMOS temperature sensor exploiting the piecewise BJT process spread compensation property of the base recombination current is proposed to reduce the process variations of the base–emitter voltage (V _be). The weighted combinations of different on-chip resistors are explored to minimise their associated process spread. Fabricated in standard 0.18-μm CMOS, the chip prototype occupies an active area of 0.036 mm² and draws 3 μA from a 1.2 V supply, with a measured maximum inter-/intra-die variation in V _be of <1.5 mV from −40 to 125°C from two batches. Using the measured V _be, ΔV _be and the first-batch-only calibration parameters, the chip prototype demonstrates an untrimmed batch-to-batch inaccuracy of ± 1.5°C (3σ) within the same temperature range (24 samples from 2 batches).

A super class AB recycling folded cascode operational transconductance amplifier is presented. It employs local common-mode feedback using two matched tuneable active resistors, allowing to adapt the amplifier to different process variations and loads. Measurement results from a test chip prototype fabricated in a 0.5 μm CMOS process validate the proposal.

As an extension of the High Efficiency Video Coding standard, Scalable High-efficiency Video Coding (SHVC) has been developed to satisfy the challenges posed by heterogeneous environments. A layer-based scheme is employed to support various forms of scalability, but this results in significantly increased computational complexity. A fast Coding Unit (CU) size decision algorithm is proposed to alleviate the additional computational requirement of the Enhancement Layers (ELs). The inter-layer dependency and the spatial correlation are jointly used to skip and early terminate the CU size decision process. In addition, the quantitive relationship between the texture homogeneity and the AC coefficient energy is investigated to further narrow the depth range to be evaluated. Experimental results show that the proposed scheme saves encoding time by up to 72% compared with the SHM 12.0 implementation, with negligible degradation in Rate-Distortion performance.

Real-time arbitrary style transfer is based on a feed-forward network, which consists of a pre-trained encoder, a feature transformer, and a trainable decoder. However, the previous approach has some degrade in style quality of output image because the pre-trained encoder is not optimised for image style transfer but originally for image classification task. An end-to-end learning scheme is introduced that optimises the encoder as well as the decoder for the task of arbitrary image style transfer. Experiments conducted with a public database proves that the style transfer network trained with the end-to-end learning scheme outperforms the network with a fixed encoder in terms of minimising both content and style losses and quality of the stylised images.

High-efficiency video coding (HEVC) is based on integer discrete cosine transforms (DCTs) of size 4 × 4, 8 × 8, 16 × 16 and 32 × 32 whose elements are coded on 8 bits. However, the algorithm requires that the output length at each processing stage should never exceed 16 bits. The conventional solution is the truncation of the least significant bits which leads to erroneous results and a waste of resources. In this Letter, different DCT kernels with reduced elements length are proposed. They are evaluated for compression efficiency and hardware implementation competence. An implementation on the Xilinx FPGA virtex-6 circuit has given a maximal operating frequency increase of 4.81 and 93.41% for the DCT-II 4 × 4 and the discrete sine transform, while reducing the energy consumption by 10.64 and 30.77% at 100 MHz, respectively. Using the HM 16.3 HEVC model and video sequences of different resolutions, the results show a quality degradation of 0.01 dB for a bit rate increase of 0.19% compared to the reference cores.

Image ordinal classification has drawn substantial attention from the research community due to the ordering relation between image categories. Recent advancements towards image ordinal classification lie in applying deep neural networks [convolutional neural network (CNN)]. Nevertheless, the lack of ordinal training data prevents deep models from generalising to testing data. In this work, two multi-view learning approaches are proposed to tackle the insufficient data issue. On one hand, a multi-view ordinal classification with multi-view max pooling (MVMP) approach is proposed, in which each image is randomly blocked with some grids thus creating multiple views of the original data. All views are then used to train multi-view CNN for classification. On the other hand, in order to account for the ordinal relation, the authors propose a double-task learning on MVMP for classification and average pooling for regression. The task of regression benefits that of classification, mainly focusing on improving classification's recognition accuracy. The two proposed approaches are validated on Adience dataset, and show very compelling results. The code and models will be available online.

The purpose of researches on multi-focus image fusion is to obtain a composed image where the objects are all captured in focus. Compared with the source images, the new one is of richer information and much better visual performance. Deep convolutional neural network (DCNN) and extreme learning machine (ELM) are combined to be a novel model (DCELM) to deal with the issue of multi-focus image fusion. First, the source images are input into DCELM. Then, ELM is responsible for generating random weights between adjacent layers. Moreover, a convolution layer followed by a pooling one forms the basic unit of DCELM, which is used to get the feature maps of the source images from different perspectives. Finally, the above features are classified via ELM, and the information in focus from the source images can be fused into the final fused image. Experimental results demonstrate that the proposed fusion method well combines the better feature extraction ability of DCNN and much faster training speed of ELM, and its performance is superior to current state-of-the-art typical ones.

A new goodness-of-fit test for spectrum sensing in cognitive radios under heavy-tailed noise is proposed, based on the geometric power (also called the zero-order statistics) of the received observations. The noise statistics is assumed to follow a symmetric-alpha-stable distribution, motivated by statistics observed in realistic scenarios. The expressions are provided for the test statistic and the asymptotic detection threshold, in terms of the number of observations under the null hypothesis. Through extensive Monte Carlo simulations, the superior performance of the proposed technique over existing non-linear detection techniques is demonstrated, such as the fractional lower-order statistics, zero-memory non-linear and myriad filtering. In addition, the advantages of the proposed technique on experiment-captured data are demonstrated.

A silicon–germanium BiCMOS E-band power amplifier (PA) for millimetre-waves for Gbit/s data rates in fifth-generation cellular wireless backhaul networks is presented. A 69–78 GHz transmission-line PA is demonstrated by integrating an electrostatic discharge (ESD) and a power detector together. To save this area, a transmission-line-based power coupler is shared by the ESD and the power detector. The ESD achieves an insertion loss of <1.5 dB in the frequency range of 17–100 GHz. The linear-in-dB output characteristic of the power detector is achieved by an exponential voltage convertor. The designed PA measured a maximum gain of 18.6 dB with 3 dB bandwidth of 68.9–78.7 GHz. The input return loss is better than −10 dB from 50 to 81 GHz. It delivers 10 dBm saturated output power at 73.4 GHz. Transmission line pulsing (TLP) measurement result shows that the voltage is clamped below 6 V even under 10 A ESD current.

Successful experimental demonstration and field trial of 400GBASE-LR8 is reported based on 4-level pulse amplitude modulation format over standard G.652 single-mode fibre using CFP8 pluggable optical transceivers. Stable long-term error-free performance of 400GBASE-LR8 optical signal over fibre spools of varying lengths up to 35 km is demonstrated, which is significantly longer than the IEEE standard 802.3bs specification of 10 km reach. For the first time, error-free full throughput performance of 400GBASE-LR8 over a 27 km field fibre link including optical patch panels through operational exchanges is also demonstrated. The results demonstrate that this early 400GBASE-LR8 client technology is suitable for practical network deployment scenarios and has potential for even longer reach beyond the 10 km specification.

To deal with a command of step load change for a dual-active-bridge converter, a transient current control with a multi-phase-shift gating scheme is needed. By selecting a reference signal with a floating position, the effective transient current control can be achieved by moving multiple voltage edges in sequence, which would fulfil the step transition of the load level in a fast and smooth manner. Although the analysis is derived using triple-phase-shift schemes, the proposed control can also be applied to other multi-phase-shift schemes, such as extended-phase-shit and dual-phase-shift scheme. An experimental test validated the correctness and effectiveness of the proposed transient current control.

Spaceborne synthetic aperture radar (SAR) polarimetric application, operating at P-band, is significantly influenced by ionospheric Faraday rotation (FR). Various researches have been published to evaluate the FR effect on polarimetric scattering matrix, nevertheless all of which are based on the narrow-band hypothesis and ignore the polarimetric dispersion (PD) effect caused by FR for its high disturbance within the bandwidth. The existent model is no longer well applicable to the future P-band ultra-wideband system with benefits of the high-resolution detection of the subsurface or obscured targets. A novel analytical model of PD based on the frequency-domain range pulse compression is proposed and analysed in this Letter, and the simulation results indicate the spaceborne P-band ultra-wideband SAR system is seriously influenced by the PD effect.

An experimental study of extreme high-squint (i.e. forward-looking) synthetic aperture radar (SAR) is presented. Experimental data collected against a distributed complex target are successfully focused using the backprojection algorithm. Interferometric SAR processing is then performed in order to overcome left-right ambiguity and estimate constituent scatterer Cartesian coordinates.

A general closed expression of posterior Cramer-Rao lower bound (PCRLB) for wireless sensor networks is derived and presented, based on Bayesian estimation and well suited to dynamic systems. The aim is to guide the localisation algorithm design and anchor layout fulfilling certain accuracy requirement. There is a growing need for wireless sensor systems to determine the position information indoors and in somewhere satellite positioning cannot provide service. PCRLB sets a fundamental lower limit to all kinds of localisation algorithms within consideration of dynamic state transition. This result shows that the theoretically best position solution can be obtained.

The enhancement in electrical characteristics of the thin film transistors (TFTs) based on the mixed-stacked amorphous zinc tin oxide (a-ZTO) with an alternating precursor concentration is reported. The channel layers were deposited via sol–gel on oxidised p-Si. The source/drain ohmic contacts were prepared on the ZTO layer, constructing the bottom-gate TFTs. In this investigation, the TFTs with the following three channel layers were fabricated, and their characteristics were compared; (i) four layers produced from 0.05 M precursor solution, (ii) four layers produced from 0.2 M precursor solution, and (iii) four layers with alternating 0.05 and 0.2 M precursor solutions. It was found that the mobility (5.3 cm²/V s) of the TFT fabricated with the channel layers with alternating precursor concentration was higher than those with single concentration. Although the mechanism for this mobility enhancement is still being further analysed, it was conjectured that alternating precursor concentration might have reduced the trap density in the transistor channel and/or ZTO/SiO₂ interfaces.

The block orthogonal matching pursuit (BOMP) algorithm is an efficient method in compressed sensing (CS) for the reconstruction of block-sparse signals, whose non-zero entries occur in clusters. However, due to the non-ideal factors in practice, there exits perturbation in the CS system, which may cause significant performance degradation during reconstruction. In this Letter, a perturbed BOMP algorithm is presented to deal with this problem, which extends BOMP algorithm to the perturbation case. The proposed algorithm performs controlled perturbation on each selected block of support vectors to decrease the orthogonal residual at each iteration. Moreover, the condition of successful reconstruction is derived. Simulation results demonstrate the effectiveness and robustness of the proposed algorithm.

In many i-vector based speaker recognition frameworks, the key challenge is to develop effective channel compensation methods for enlarging inter-class differences while reducing intra-class variations. This challenge is handled with a discriminatively learned network (DLN), which uses both speaker classification and verification signals as supervision. The speaker classification task forces the embeddings (vectors mapped from i-vectors) of different identities drawing apart to increase the inter-class variation, while the verification task pulls the embeddings of the same identity together to reduce the intra-class variation. DLN projects i-vectors to a more discriminative embedding space. However, the verification scores are cosine similarities between these embeddings. The learned DLN can be well generalised to new speakers unseen in the training data. On the text-dependent challenging Robust Speaker Recognition (RSR2015) database, the performance is significantly improved when compared with the linear discriminant analysis (LDA) and Gaussian probabilistic LDAmethods.