In coplanar waveguide (CPW) circuits, printed balanced antennas must be excited through baluns. These baluns often feature coplanar-to-slotline transitions that must be tuned and matched by electromagnetic optimisation, and may interfere with the antenna backward radiation. A new CPW balun for printed balanced antennas is presented and modelled. As its multimodal circuit model makes clear, it integrates an impedance-matching network within its structure that can be analytically designed. The balun has been experimentally tested, exhibiting good agreement with its circuit model and a weak electromagnetic interaction with the antenna radiation.

A reconfigurable dual-port diversity antenna for ultra-wideband application is presented. The proposed design can be configured either in stacked/back-to-back arrangement by placing radiating elements facing in the opposite directions or in orthogonal arrangement. The proposed designs are realised on FR4 substrate of dimensions 40 × 36 × 1.5 mm. By introducing an efficient decoupling structure in the ground plane, both configurations maintain an isolation better than 25 dB over the complete band. Performance of the proposed design is verified by the measured results including s-parameters, envelope correlation coefficient, total active reflection coefficient and channel capacity loss. The proposed design is suitable for integration in three-dimensional (3D) system-in-package applications.

A novel concept of frequency-independent variable phase shifters is introduced. It is based on the use of circularly polarised (CP) antennas coupled in the near-field. For transmission between two CP antennas in the far-field, a rotation around the boresight axis produces a phase change in the transfer function. This concept is exploited and validated also for near-field coupling between two CP antennas. Therefore, two close-coupled CP antennas can act as an adjustable phase shifter with the phase shift controlled by the angle between the structures. The bandwidth depends on the CP bandwidth of the antenna. This concept is valuable for future phase shifters in millimetre-wave applications where rotation can be facilitated by MEMS elements.

Magneto-acoustic (MA) tomography combines near-field RF and ultrasound for the goal of a safe, high-resolution, high-contrast bimodal imaging technique. Coherency between RF excitation and the resulting ultrasound generation through Lorentz force interactions is established. This allows for continuous-wave (CW) imaging techniques to replace conventional pulsed-based approaches, easing system design and improving signal-to-noise ratio (SNR) while reducing the peak-to-average excitation power ratio, for the same averaging time, sensitivity and gain. This is the first successful demonstration of MA detection using step frequency CW (SFCW) as well as frequency-modulated CW (FMCW) approaches with up to 36 dB SNR improvement.

An all-analogue feedback duty-cycle corrector (DCC) circuit with high precision and frequency is presented to tighten duty cycle into an allowable range and compensate for duty-cycle uncertainties in high-speed interfaces. The proposed DCC is employed to calibrate the duty cycle of the clock to reduce the deterministic jitter introduced by the duty-cycle distortion. It extracts the duty-cycle information by a differential duty amplifier detection scheme and corrects the clock distortion by a duty-cycle adjuster through the negative feedback loop. The DCC has improved robustness, correction range and operating frequency as compared with other DCCs. With post-simulated results using 55 nm CMOS technology, the output duty cycle is corrected to 50 ± 0.1% over the input duty-cycle range of 20–80% for 1–5 GHz. It consumes 3.6 mW at 3 GHz using a 1.2 V supply voltage and occupies an area of only 0.00174 mm².

A novel double-sampled wideband delta–sigma modulator topology with shifted loop delays is proposed. Compared with the conventional double-sampled modulator, this analogue-to-digital converter (ADC) implements the inherent quantisation delay by shifting one loop delay from the last integrator to the quantiser, and it relaxes the critical timing for dynamic element matching (DEM) by shifting the loop delay from the first integrator to the feedback path. In addition, by inserting one more delay in the signal path, the proposed modulator retains the low-distortion property. To verify the effectiveness of the proposed topology, a second-order double-sampled delta–sigma modulator is analysed and simulated.

A simple and uniform compensated bias for a fixed-width squarer is proposed. By employing the expected value of truncation parts, truncation errors can be efficiently alleviated. The proposed squarer derives a simple compensated bias and achieves a small area. The results show that the proposed squarer achieves low-cost, high-accuracy performance and yields higher values than previous designs do, particularly regarding the value of accuracy divided by the area–delay product.

A new method for time-domain sensitivity analysis of on-chip coupled interconnects with non-uniform substrate temperature profiles is presented. The parametric macromodelling is performed via multivariate rational approximation using magnitude vector fitting algorithms. The waveform relaxation algorithm with transverse partitioning is employed for efficient sensitivity analysis of coupled interconnects. With these techniques, the thermal sensitivity of on-chip global interconnects arising from typical substrate thermal profiles are investigated.

Chemical solution deposition (CSD)-derived ferroelectric PbLaZrTiOx (PLZT) capacitors with an aluminium-doped zinc oxide (AZO) top electrode is fabricated. The effects of a thin conductive AZO buffer layer between the Pt bottom electrode and PLZT were investigated. The hydrogen degradation resistance of PLZT capacitors with a 10 and a 20 nm AZO buffer layer was improved up to 91 and 81%, respectively, of an initial polarisation value even after 45 min annealing in a 3% hydrogen atmosphere, from 42% of that without a buffer layer. The fatigue endurance (at 200 kV/cm (10 V) at 100 μs pulse width at 1 ms intervals) of the PLZT capacitor with a 10 nm AZO buffer layer was also improved after 10⁷ cycles.

Photo-induced coded-aperture imaging (PI-CAI) based on compressed sensing (CS) at 590 GHz using a WR-1.5 (500–750 GHz) vector network analyser is demonstrated. For a 256-pixel (16 × 16) frame, the acquisition time can be reduced by 40%, as compared with PI-CAI without CS, while maintaining high imaging quality. On the basis of this approach, it is envisioned that real-time (26 fps) THz imaging with 1000 pixels (32 × 32) can be realised using high-speed digital micromirror device chipsets and optimised data-acquisition software.

A new approach regarding real-time low-level image and signal processing is presented. This new approach deals with the philosophy of finding a method for designing and executing algorithms (filters), using the most minimum possible complexity. The proposed method is based on a special formation of coordinate logic filters (CLFs), the coordinate logic order statistics (CL-OS) filters. CL-OS introduces a minimalistic approach in filter design followed by the greatest possible efficiency in terms of speed, quality, complexity and energy consumption. Moreover CL-OS reveals new signal information; this information gives a different interpretation of the original signal, it could be said that it acts like a function applying a nonlinear quantisation in order to approximate the desired output.

High-efficiency video coding (HEVC) is the state-of-the-art video coding standard, which adopts more complicated and time-consuming intra-prediction (IP) modes. Many-core processors are good candidates for speeding up HEVC IP in the case that HEVC IP can provide sufficient parallelism. Proposed is an efficient parallel framework for HEVC IP. Experiments show that the proposed method dramatically accelerates more than the state-of-the-art parallel method.

Brute force matching of binary image feature descriptors is conventionally performed using the Hamming distance. The use of alternative metrics is assessed in order to see whether they can produce feature correspondences that yield more accurate homography matrices. Two statistical tests, namely analysis of variance (ANOVA) and McNemar's test were employed for the evaluation. Results show that Jackard-Needham and Dice metrics can display better performance for some descriptors. Yet, these performance differences were not found to be statistically significant.

Radial symmetry detection is a hot topic in computer vision owing to its importance in understanding the image structure and content. A major generalisation to the problem is presented and single image spiral symmetric centroid detection is studied. Since radial symmetry is just a particular case in the model of spiral symmetry, spiral symmetry detection is more challenging and general. For this new problem, a fast and effective voting-based solution is proposed, which can reliably detect both the classical radial symmetric centre and the more general spiral symmetric centroid. Extensive experiments validate the generality and practicability of the proposed approach.

Complementary sequences are being widely used in applications ranging from communication to radar and sonar systems. The properties of these sequences render them suitable for applications requiring signal detection in noisy environments and/or multi-user systems. The use of complementary sequences in practical applications with reduced hardware resources has promoted the development of efficient architectures that reduce the computational requirements for the generation and correlation of the sequences. An optimised architecture for the simultaneous correlation of pairs of multilevel complementary sequences is presented, which achieves that reduction by removing redundant operations. The proposal has an impact also on the correlation of binary complementary pairs of sequences of lengths of a power of 10 and 26 that use the multilevel architecture to generate the kernels.

A 10.2–12.6 GHz balanced heterojunction bipolar transistor (HBT)–high electron mobility transistor (HEMT) frequency tripler is presented. A pair of common-base/common-emitter HBTs is used to generate in-phase and out-of-phase harmonics. Two bandpass filters are utilised as matching networks to enhance the harmonic suppression. A common-gate/common-source HEMT active balun is employed to combine the third harmonic in-phase and provide conversion gain. The proposed frequency tripler shows a conversion gain of 2.8 dB, a fractional bandwidth of 21.2% and a fundamental suppression higher than 47 dB.

A 60 GHz upconversion mixer is realised in a 65 nm LP CMOS. The mixer consists of a Gilbert cell, a LO buffer and a RF buffer. To enhance the bandwidth and reduce the area, the transformer impedance matching network is used. The measured results show that the mixer covers the IEEE 802.15.3c standard four channels, from 57 to 66 GHz. In each channel, the gain flatness is <1.5 dB. In channel 2, the mixer achieves a gain of 14 dB and an output power 1 dB compression point of 2.5 dBm, while consuming 32 mW power drain from a 1 V supply. To the best of authors’ knowledge, this is an advance in the state-of-the-art mixers.

A filtering power divider (PD) with good isolation performance is presented. The isolation performance is improved by introducing a distributed isolation network (DINW) to replace the conventional resistor isolation network. Moreover, the proposed DINW can be incorporated into a filter design to form a filtering PD. The filtering power divider was implemented at a centre frequency of 1 GHz with a fractional bandwidth of 56%, and the measured isolation is better than 23 dB across the entire operating band.

The first ever demonstration of a viable contactless radio frequency (RF) probing technique at elevated temperatures is presented. The design utilises an inverted microstrip design for the probe, which is suspended over and placed in close proximity to the input/output microstrip lines of the device under test to couple signals in and out. To demonstrate the efficacy of the contactless RF probing technique, three example circuits, namely, a microstrip spurline bandstop filter, a microstrip pin diode series switch and a monolithic microwave integrated circuit amplifier mounted on a microstrip line were designed, fabricated and performance measured against temperature up to 200°C.

A method to determine the capacitors that represent the periodical loads of transmission lines embedded between woven fibre substrate layers is presented. The resulting model allows an accurate modelling of the transmission characteristics of striplines presenting double resonances when operated at microwave frequencies.

A temperature-based tunable dielectric resonator is presented showing a tunability of 20 MHz around 11.8 GHz when heated with a drop of the quality factor from 4000 to 286. The puck-shaped alumina resonator contains a thin sapphire sheet coated with VO₂. An overview of the thermoconductive VO₂ material used is presented. The design and simulations of the resonator are presented along with the fabricated structure and measurements.

The highly stable operation of 980 nm transverse coupled cavity vertical cavity surface emitting lasers (VCSELs) with enhanced modulation bandwidth for high-speed optical interconnects is reported. The device operates at 25 Gbit/s from 0 to 60°C with fixed driving conditions. In addition, the stability of driving currents into the VCSEL is explored, exhibiting that the bandwidth enhancement is not sensitive to the current. The ultra-compact inplane coupled cavity VCSEL can boost the modulation speed beyond the relaxation oscillation frequency because of the to photon–photon resonance, which makes it useful in next-generation computing and data centre networking.

Fault detection and identification are new and challenging tasks for electrical generation plants that are based on solid oxide fuel cells. The use of a quantitative model of the plant together with a support vector machine to design and operate a supervised classification system is proposed. This type of system, which uses a few easy-to-measure features selected through the maximisation of a classification error bound, proved to be effective in revealing a faulty condition and identifying it among the four considered fault classes.

Directly interfacing sensors that link sensors to micro-controllers without a signal conditioning circuit or an analogue-to-digital converter on the one side and energy harvesting systems on the other side are two key enabling technologies to create self-powered autonomous low-cost and low-maintenance sensor networks. In this reported work, both technologies are concurrently combined in such a way that the duty cycle of the pulsed power supply provided by the harvesting system itself yields in turn a temperature gradient sensor functionality. Originally conceived to power a low-power satellite beacon in a CubeSat project, it has found application, among other applications, in automatic air conditioning and heating systems aiming smart buildings.

A wireless indoor positioning system using white LED lights is proposed. The time difference of arrival technique is employed and the phase differences between the received signals are determined to develop a positioning algorithm which can estimate the receiver location with a mean localisation error as low as 1 mm in a room of dimensions 5 × 5 × 3 m. Through simulations, it is identified that the optimum receiver height where localisation error gets minimised is between 2.5 and 3 m from the ceiling which corresponds well with the typical dimensions of a room.

To focus synthetic aperture radar data suffering from phase errors, a novel autofocus algorithm is presented. The algorithm uses the coordinate descent technique to estimate the phase errors for maximum sharpness autofocus, where analytical-form expressions are easy to find for the optimal estimates on a per-parameter basis. Experimental results show the validity of the proposed algorithm.

The main objective of compact device modelling is to perform the best possible fitting of the model to the measured characteristics. To do so, there are some parameters in the models that must be adjusted to the actual values used in the technology, plus some other parameters that are used to overcome some of the limitations of the assumptions used in the model development. Obtaining the first set is usually straightforward, whereas determining the second set is more difficult, because the extraction methods are not always efficient and, moreover, there are parameters that are correlated with others, complicating the whole issue. Moreover, there may be parameters that are not relevant for the specific technology. If one could be able to determine the importance of different parameters, then effort could be better directed by dedicating more effort to the most important parameters. A method to order a list of given model parameters using the calculated importance as the criterion is proposed. This importance is calculated using a modification of the Google PageRank algorithm.

A novel computational technique is presented to study the terminal influence inside the three-dimensional (3D) nano-scale metal-oxide semiconductor field effect transistor (MOSFET) using TCAD simulations. Within the MOSFET the derivative of the electrostatic potential with respect to voltages at each terminal is taken, and when these derivatives are summed together they always sum to unity. It is found that these functions can be used to quantify the relative influence or control of the terminals anywhere inside the MOSFET, including the channel. The motivation for moving from planar MOSFETs to 3D-MOSFETs is to increase the gate control over the channel. The terminal influence functions quantify the notion of control. To gain insight into the working of a semiconductor device, different quantities like potential, charge or current density etc. may be visualised. These quantities are available in the standard TCAD tool-kit. However, these do not directly address the mechanism of terminal control. The terminal response or control functions can be used to do this very clearly.

A pseudo-cascode split-transistor technique is proposed for DC-gain enhancement of amplifiers in the subthreshold by exploiting the small size effects of metal–oxide semiconductor field effect transistors (MOSFETs) including the reverse short-channel effect and the inverse narrow-width effect. It requires no body-biasing and occupies a small area. A compact 114 μm² two-stage amplifier with pseudo-cascode compensation for in-pixel amplification in vision sensors has been designed using the proposed technique. A total of 10 samples of split-transistors and amplifiers fabricated in an UMC 0.18 μm standard CMOS process were measured. More than half of the tested split-transistors show considerable DC-gain enhancement over a wide range of bias currents and nine amplifiers have increased DC gains larger than 85 dB at about 4 nA power consumption.

Formulated in the complex-domain, an exponential integral expression is presented for the generalised Marcum Q-function. Specifically, the generalised Marcum Q-function is represented as a finite-range single integral whose integrand is exponential in both parameters of the generalised Marcum Q-function. Consequently, this exponential representation leads to a unified moment-generating function (MGF)-based approach to performance analysis of energy detection in general fading channels. The MGF approach is generally applicable for energy detection with or without diversity in arbitrarily (maybe correlated) fading channels, provided that the MGF of the received signal-to-noise ratio is available.

A novel wide-pipeline low-density parity-check (LDPC) decoder approach for the worldwide interoperability for microwave access (WiMAX) standard (802.16e) is proposed for execution on field-programmable gate arrays (FPGAs), using a high-level synthesis tool to reduce the development effort and design validation time that generates a wide-pipeline architecture. Optimised open computing language (OpenCL)-based kernels are developed and the integration of distinct configurations of single instruction multiple data and compute units to increase the level of parallelism are analysed. The decoding throughput surpasses the minimal requirements of 75 Mbit/s, a key figure of merit that ranks the design with other very large-scale integration-based approaches. Furthermore, extra precision is deployed with 8-bit fixed-point arithmetic, delivering superior bit error rate performance and lower error floor regions.