A novel bandwidth-enhanced UWB tapered slot antenna with Y-shaped corrugated edges is proposed. In the double-slot structure, the two slots are separated by a V-shaped metal surface with straight edges, which is beneficial to improve the directivity of the antenna. Meanwhile, an exponential Y-shaped corrugated edge is designed. This novel corrugated edge not only enhances the impedance bandwidth, but also improves the gain by concentrating the energy near the tapered slot. The impedance bandwidth of the double-slot TSA is from 2.5 to 28 GHz. The gain of the antenna is >10 dB from 3.5 to 25 GHz, and >8 dB in the whole operating band.

A loop-type ground radiation antenna is proposed for a discontinuous square ring-shaped ground plane. The antenna shows strong coupling with the ground mode if it is located at the maximum current density location. The location of the antenna has been evaluated using characteristic mode analysis of the currents on the ground plane. The size of the loop-type ground radiation antenna is 10 × 4.5 mm and it is fabricated on a square ring-shaped ground plane of size 50 × 50 mm having a gap of 2 mm. The design is intended for mobile device applications operating at 2.45 GHz.

Rapid calculation of electromagnetic (EM) scattering problems over a wide incident angle is always a difficult task for computational electromagnetics (CEMs). As a new kind of excitations which include plenty of information about incident angles was constructed, a fast method by applying compressed sensing in conjugation with traditional method of moments has been proposed. Random sampling is firstly taken into the process of linear measurement to simplify the construction of the new excitations and accelerate the recovery of unknown induced currents. The basic principle is elaborated in detail, and the effectiveness of the proposed method is validated with numerical results.

In this Letter, a V-band chip-level dual-polarised dielectric resonator antenna implemented by using bondwires and silicon-based integrated passive device technology is proposed. The square-shaped resonator is fed by two bondwire coupling structures which excite two degenerate modes orthogonal to each other. The resonance of bondwire itself is also found to enhance the antenna bandwidth to cover the 60 GHz band. Reasonable agreement between the simulation and the measurement is obtained. The measured antenna bandwidth is from 52.8 to 65 GHz. The measured isolation is better than 20 dB at frequencies of interest. The measured antenna gain is 4.5 dBi at 60 GHz. The proposed design can provide further applications for circularly polarised systems.

A 266–2133 MHz phase shifter is proposed for LPDDR4X interface, utilising an all-digital delay-locked loop (DLL) and a triangular-modulated phase interpolator (PI) to improve the jitter and linearity. The DLL consists of two kinds of DLLs: a global DLL to assist fast locking; and a local DLL which uses an adaptive-window phase detector and a folded delay line to reduce jitter and improve linearity. The PI uses a triangular-modulated clock waveform to achieve good linearity over a wide frequency range. The prototype chip is implemented in a 65 nm CMOS process. The measured jitter of the DLL is 3.08 ps_rms/19.93 ps_pp at 2133 MHz. The measured differential non-linearity of the phase shifter is <0.91 LSB at 2133 MHz, and the power efficiency is about 2.7 mW/GHz.

To address the conflicting requirement of fast convergence rate and low misadjustment, a new non-parametric variable step-size diffusion least mean squares (LMS) (NPVSSDifLMS) algorithm is presented. The non-parametric variable step-size scheme is obtained to cancel the influence of noise signals in the intermediate posteriori errors at every node. Simulation results show that this NPVSSDifLMS algorithm outperforms the DifLMS and VSSDifLMS algorithms.

Correlated double sampling (CDS) circuits are essential to process the X-ray charge-coupled devices (CCDs) that have been used in the modern X-ray astronomical field. A novel CDS circuit is presented. The innovation of the circuit is that it greatly improves the noise and linearity performances through moving the clamping switch out of the signal path as well as omitting the sampling switch which eliminates the nonlinearity and thermal noise introduced by these switches. Also, a baseline adjustment (BLA) circuit is incorporated so as to make efficient use of the dynamic range of the ADC, making the CDS circuit flexible to interface the CCD operating under a wide range of temperatures. A prototype application-specific integrated circuit (ASIC) is designed with GlobalFoundries 0.35 μm CMOS process and the experimental results show that with pixel rate varying from 100 kHz to 1 MHz, the integral nonlinearity and equivalent input noise performances are quite consistent, which are better than 27 ppm and 23 μV, respectively, within a dynamic range of 1.5 V_pp. The effective adjustment range of the BLA is at least 2 V, and the power consumption for a single CDS core is 10 mW approximately from a 3.3 V power supply.

A power-efficient complementary voltage-to-time converter (CVTC) is proposed for a flash ADCs. The alternating reset direction according to the signal development direction reduces the power consumed by the reset operation and the operational frequency of the CVTC is effectively reduced by half. Accordingly, the logic circuits following the CVTC work in a time-interleaved manner, resulting in significant power saving. A 5-bit 2.5 GS/s flash ADC designed for a 40 nm CMOS process shows a 27% power reduction over the conventional voltage-to-time conversion-based flash ADC.

A statistical estimator based on maximum-likelihood estimation theory is developed to simultaneously reduce capacitor mismatch and noise in a successive approximation register analogue-to-digital converter (SAR ADC). After the SAR ADC has finished quantisation, the residue voltage is available at the comparator input and is estimated accurately by using the statistical estimator. The ADC resolution is improved by subtracting the estimated residue from the digital output. The same technique of residue extraction is used to estimate mismatches in the capacitive digital-to-analogue converter. A 7 dB improvement is shown in signal-to-noise-plus-distortion ratio by using the statistical estimator for an 11-bit SAR over a wide range of capacitance mismatch and ADC noise.

This Letter proposes a new two-sample excess loop delay compensation in continuous-time delta-sigma modulators using time-shifted differentiated digital-to-analogue converter (DAC). This is done by using the information inherently available but ignored in the conventional methods, and therefore no additional hardware is required. Simulation results and mathematical analysis are provided to verify the effectiveness of the proposed structure.

A hybrid modified teaching–learning-based particle swarm optimisation (HMTL-PSO) initialised by the normalised step cost (NSC), named HMTL-NSCPSO, is proposed for solving trajectory optimisation with complex constraint problems. Specially, the new HMTL-NSCPSO combines the canonical PSO basic policy, the teaching–learning-based optimisation (TLBO) algorithm and the normalised step cost (NSC) function in order to promote diversity, obtain well-speed convergence and to improve search ability. The algorithm is tested on an UAV trajectory optimization problems. Experimental results validate the effectiveness of the HMTL-NSCPSO.

For the first time, the design, fabrication and testing of a dispenser-printed electrochromic (EC) display on fabric using Poly(3,4-ethylenedioxythiophene)-poly(styrenesulfonate) (PEDOT:PSS) as the EC material are reported. The entire display is based on a lateral structural design in which the electrodes are located adjacent to each other in the same plane. The display is directly printed onto a, polyvinyl chloride (PVC)-coated, 100% polyester woven fabric which is commonly used in the creative industries. Each layer was cured below 130°C to ensure no thermal damage to the fabric. Each display pixel, consisting of the colour changing and counter electrodes, was separately driven at two voltage direct currents (VDCs). The colour change between pale blue and dark blue was controlled by switching the polarity using a microcontroller. Two demonstrators, a 3 × 3 pixel matrix display and a seven segment display, were achieved with an average switching speed of 5 s.

A method for the adaptive signal decomposition by using the frequency boundary information is proposed. This is achieved by the high-order derivative operation that could extract the frequency-domain boundary between the highest frequency portion and the rest parts of signal by using the proper order value and a subsection filtering method based on fast Fourier transform. Simulations show that the proposed method performs better in frequency resolution, modal aliasing than empirical mode decomposition, ensemble empirical mode decomposition (EEMD) and complementary ensemble empirical mode decomposition (CEEMD).

The importance sampling (IS) method lies at the core of many Monte Carlo-based techniques. IS allows the approximation of a target probability distribution by drawing samples from a proposal (or importance) distribution, different from the target, and computing importance weights (IWs) that account for the discrepancy between these two distributions. The main drawback of IS schemes is the degeneracy of the IWs, which significantly reduces the efficiency of the method. It has been recently proposed to use transformed IWs (TIWs) to alleviate the degeneracy problem in the context of population Monte Carlo, which is an iterative version of IS. However, the effectiveness of this technique for standard IS is yet to be investigated. The performance of IS when using TIWs is numerically assessed, and showed that the method can attain robustness to weight degeneracy thanks to a bias/variance trade-off.

This Letter proposes a miniaturised filter that makes use of quarter-mode substrate-integrated waveguide (QMSIW) combining with a comb-line architecture for the first time. The comb-line QMSIW filter uses a periodic structure of distributed capacitance thereby reducing the size of the conventional QMSIW filter. To enhance out-of-band rejection, an S-shaped slot is introduced to realise negative coupling. Using the above principles, the size of this filter is reduced by nearly half (compared with conventional QMSIW filters) while having high out-of-band rejection. Finally, this planar, miniaturised filter is designed and fabricated. The measured results match well with simulated results.

A miniaturised balun filter with tunable power division ratio (PDR) and centre frequency is presented. Compared with the conventional balun filter, the proposed one with a varactor diode and a slide rheostat loaded onto transmission line (TL) reduces the electrical length of TL from 180° to 108°. Meanwhile, the varactor diode and slide rheostat are also utilised to control the PDR at two output ports. Moreover, passband frequency can be tuned by controlling the varactor diodes loaded onto the dual-mode resonators. The proposed balun filter achieves a tunable power division from 1:1 to 3:1 and 28.6% wide frequency turning with better than 12 dB return loss. The total electrical size of fabricated circuit is 0.23 × 0.12λ_g .

In this Letter, a flat terahertz (THz) transmit array with a two-layer element is investigated. The proposed element can provide 360° phase coverage with a transmission coefficient better than −3.52 dB at 250 GHz. Then a transmit array based on the element is designed with a low focal-to-diameter ratio of 0.3. Simulation results reveal that the transmit array has large backward radiation. A WR-03 open-ended waveguide with a thick wall is used to feed the transmit array, thus the backward radiation is further reflected to the forward direction and the gain is increased by 1.7 dB at 250 GHz. Then the transmit array is fabricated using standard, low-cost PCB technology, and the measured results show that a peak gain of 28.8 dBi with an aperture efficiency of 32% is achieved. In general, the proposed transmit array has simple structures and can be fabricated by standard PCB technologies which are suitable for applications in low-cost THz systems.

A compact filtering power divider composed of two quarter-mode circular cavities based on the substrate integrated waveguide is proposed. Two extra perturbing via-holes added in each side of the quarter-mode circular cavity achieve a selective bandpass filtering response and improve the performance of the upper stopband. The proposed FPD operates at 9.1 GHz with a 3 dB fractional bandwidth of 19.8%. The insertion loss in operating band is < 1.6 dB and the input return loss is better than 16 dB. The amplitude and phase imbalances between the two outputs are measured better than 0.6 dB and 3.5°, respectively. The measured results exhibit good agreement with the simulated ones.

The steady-state mean-square deviation (MSD) analysis of the sign subband adaptive filter is proposed. The proposed analysis is derived by the update recursion of the MSD using Price's theorem and chi-distribution in stationary environments. Experimental results show that the proposed analysis is very close to simulated results not only for white input but also for coloured input signals.

A compact wideband bandpass filter at the centre frequency of 5.77 GHz with 78% fractional bandwidth is presented. An especial coupling structure along with a novel triple mode spiral resonator designed based on modal analysis is used to provide a flat wide passband with excellent skirt characteristics. To verify the efficiency of the filter, a prototype is fabricated, tested and compared with some of the most recent works.

A compact (200 μm) spot-size converter (SSC) with a vertically tapered core layer is developed, which is fabricated with a simple wafer process using the micro-loading effect. It is confirmed that the vertical spot size of the SSC exceeded 2 μm, which was double that of the original waveguide. As a result, an InP waveguide integrated with the SSCs exhibited a coupling loss of <1 dB per facet for a lensed fibre with a 2.5 μm mode-field diameter. The fluctuation in the total of coupling and excess losses of the SSC was <0.5 dB.

Multi-Watt high-power terahertz (THz) frequency quantum cascade lasers are demonstrated, based on a single, epitaxially grown, 24-μm-thick active region embedded into a surface-plasmon waveguide. The devices emit in pulsed mode at a frequency of ∼4.4 THz and have a maximum operating temperature of 132 K. The maximum measurable emitted powers from a single facet are ∼2.4 W at 10 K and ∼1.8 W at 77 K, with no correction being made for the optical collection efficiency of the apparatus, or absorption by the cryostat polyethylene window.

A fully-integrated 12 kHz relaxation oscillator for energy harvesting application is presented. The oscillator has no comparator and no reference voltage. A conventional comparator has been replaced by a pseudo-inverter chain to avoid comparator offset effect. Moreover, inverter threshold variation has no impact on the frequency. The circuit has been designed with 0.18 µm standard CMOS process. The oscillator operates in subthreshold region and consumes an ultra-low power of 120 nW. A temperature coefficient of 31 ppm/°C has been achieved over −40 to 90°C.

A reduced look-up-tables (LUTs) implementation of digital predistortion (DPD) for concurrent dual-band power amplifiers (PAs) is proposed. In the proposed method, the dual-band output is captured jointly with reduced sampling rate, then the PA characteristics of the lower band and the upper band are identified with a group of shared model coefficients, thus DPD can be implemented with shared LUTs for the two bands, which results in reducing the LUTs by 50% comparing with those traditional dual-band DPD techniques implemented with individual LUTs for each band. Experimental results show that the proposed method can achieve better than −48.2 dBc adjacent channel power ration performance with reduced LUTs.

Single-ended resonant inverter topology is widely used in domestic induction heating (IH) application due to its concise structure and cost effectiveness. Whereas high resonant voltage caused by the energy transfer between load inductor and resonant capacitor in the IH system may negatively affect the operation of system and long-term reliability. Generally, 1200 V rating IGBT is popularly used for single-ended IH cookers. However, abnormal conditions of AC supply suddenly increases the resonant voltage thereby IGBT can be destructed. Resonant voltage limiting technology is proposed by restricting the maximum current flow to the load inductor in order to relieve the voltage stress across the IGBT. The effectiveness of the proposed method is verified for AC supply distortion by using 1.35 kW single-ended IH system.

This Letter proposes a high-precision and non-iterative polarimetric synthetic aperture radar (SAR) (PolSAR) calibration algorithm. By solving a system of non-linear equations, the distortion parameters can be obtained directly. The proposed algorithm is assessed by the simulated PolSAR data. The experimental results show that the algorithm can accurately solve all distortion parameters even though the crosstalk is high and has a better estimation performance compared with the widely used non-iterative Quegan algorithm.

By using the effects of non-ideal oscillators, a technique is proposed to discriminate targets in continuous wave (CW) tracking radar. This method is based on the fact that each oscillator has an individual effect on the phase noise that helps to recognise real targets from the simulated ones. A simulated fake target, made through complex circuits, e.g. a digital RF memory, has a different phase noise from a real backscatter. Therefore, the transmitted and received signals are different in phase noise levels and patterns. To evaluate the performance of the proposed technique, the probability of detection (P _D) is shown in different signal-to-noise ratios, ranges, and signal-to-jammer ratios based on an experimental setup.

A new process that can improve die shift issue in process utilising back-sided under bump metallurgy (UBM) via a self-alignment effect for fan-out package applications is proposed. The die and the pad of the substrate are spontaneously aligned and the degree of accuracy is very high after reflow. For the experiment, two kinds of UBM pads were prepared on glass dies and the SAC305 solder was formed on the copper pads of an frame retardent (FR)-based substrate. The die shift value was from 20 to 50 μm after pick-and-placement. After reflow, the maximum die shift value was <1 μm. An initially misaligned die with a large shift was aligned with a high degree of accuracy during reflow due to the surface tension of the molten solder. The proposed process improves the die shift issue in the fabrication of fan-out wafer-level packaging and the active die embedded substrate.

A novel phase gradient method for detecting orbital angular momentum modes is proposed. By utilising a two-element time-modulated array, the phase gradient of the received signal is estimated through analysing the modulated harmonic component. The bit error rate performances are evaluated using Monte-Carlo simulation. Results show that the sensitivity strongly depends on the number of sampling modulation pulses. Compared with the conventional phase gradient receiver, the proposed method obviously enhances the performance of detection.

Blackspots are areas of poor signal coverage or service delivery that leads to customer complaints and loss in business revenue. Understanding their spatial–temporal patterns at a high resolution is important for interventions. Conventional methods such as customer helplines, drive-by testing, and network analysis tools often lack the real-time capability and spatial accuracy required. The potential of utilising geo-tagged Twitter data to uncover blackspots is investigated. Lexicon and machine-learning natural language processing techniques are applied to over 1.4 million Tweets in London to uncover blackspots for both pre-4G (2012) and post-4G (2016) roll out. It was found that long-term poor signal complaints make up the majority of complaints (86%) pre-4G roll out, but short-term network failure was responsible for most complaints (66%) post-4G roll out.