A dual-band dual-mode patch antenna for on–on–off wireless body area network (WBAN) applications is proposed. The antenna consists of a half-circular patch with two shorting posts to generate a TM₄₁ higher-order resonant mode and a bottom ground slot to generate additional resonance. To reduce the antenna size without seriously affecting the radiation pattern, the antenna is cut in half and operates in half-mode. The antenna has a vertical monopole-like radiation pattern in the 2.45 GHz industrial, scientific and medical (ISM) band and a horizontal monopole-like radiation pattern in the 5.8 GHz ISM band. To evaluate the body effect, both 2.45 and 5.8 GHz muscle-equivalent phantoms were used. The overall dimensions of the proposed antenna is 30.5 × 62 × 3.15 mm.

A very simple circularly polarised (CP) printed patch antenna with enhanced bandwidth has been designed, fabricated and tested. The antenna consists of two semi-circular patches and a rectangular ground plane, and is excited by a simple microstrip feedline. The total substrate size is 50 × 50 × 1.52 mm³. The measured −10 dB impedance bandwidth is 101% (2.84–8.7 GHz) and the measured 3 dB axial-ratio (AR) bandwidth is 49.8% (3.95–6.57 GHz). Good agreement is achieved between measured and simulated results.

Radio-frequency (RF) induced heating at the electrodes of the deep brain stimulation lead is a major concern when used in magnetic resonance imaging (MRI). A novel lead design is proposed to address this problem. By adopting a coil with non-uniform diameters, the maximal temperature rise measured in a standard phantom is reduced from 4.6 to 2.0°C, with an average reduction of 59%. The suppression of RF heating is probably achieved because of the alteration of the transmission line parameters of the lead by the novel design. The details of the design, experiment validation and discussion of the mechanisms are presented.

A high-voltage (HV) switch is proposed which adopts the floating control circuit to reduce area by removing the passive components and reducing the number of laterally diffused metal–oxide semiconductor field effects (LDMOSFETs). The proposed HV switch consists of three LDMOSFETs and three MOSFETs without any passive components. In addition, the routing complexity of the control signals is greatly reduced by sharing all control signals in the array of switches. The proposed HV switch is fabricated using a 0.18 μm CMOS process with 50 V devices and occupies 85 × 65 μm². The measured off-isolation is −53 dB.

Branch retinal vein occlusion (BRVO) is one of the most common retinal diseases. Without timely diagnosis and treatment, it would seriously impair the patient's vision. Automatic recognition of BRVO could significantly improve the efficiency of diagnosis. A feature representation method is proposed for the automatic recognition of BRVO with fluorescein angiography (FA) images. The proposed feature representation method, termed hierarchical local binary pattern (HLBP), is comprised of LBPs in a hierarchical fashion with max-pooling. A FA image dataset is established to evaluate the performance of the HLBP method. Experimental results demonstrate the superior performance of the proposed HLBP method for BRVO recognition with FA images, by comparing it with state-of-the-art methods.

A real-time lossless compression technique for ECG signals, which benefits wearable medical devices with stringent low-power requirements, is presented. The real-time ECG waveform is automatically classified into four regions according to its fluctuation features and the most suitable prediction method is adaptively selected from several linear prediction methods for different regions. Further proposed is the use of a modified variable length code to encode the prediction difference for a simpler transmit format. Experimental results based on three publically available test databases show that the proposed method achieves a better compression ratio with a lower prediction difference than existing state-of-the-art approaches. A very large-scale integration implementation is also demonstrated which can be used as an intellectual property core with a core area of 25 809 μm² and which achieves a power consumption of 127 μW at 100 MHz in a 0.18 μm CMOS technology.

A fully-integrated low-power and area-efficient bio-potential amplifier for high-density neural recording applications is presented. Using a self-biased amplifier structure, the bio-potential amplifier can operate at a lower power with a reduced active area. The Miller capacitance and area-efficient pad techniques harnessed in the amplifier also help to dramatically shrink the area of the amplifier. This implantable bio-potential amplifier was fabricated using a 0.35 µm CMOS process and was verified on a bench top by observing ex vivo biological signals.

A high-order curvature-compensated current mode bandgap reference (BGR) over a very wide temperature range is presented. High-order curvature correction for this reference is accomplished by the proposed multipoint corrected technique, which realises exponential curvature-compensation terms in lower and higher temperature ranges separately through simple structures. The compact multipoint curvature-compensation circuit cancels out the nonlinear terms of the BGR using sub-threshold operated metal–oxide semiconductor field effect transistors, which subtract those currents that are proportional to the nonlinearity out of the reference voltage. As a result, a flattened and better effect of curvature compensation in a wide temperature range is realised. The circuit performance was verified experimentally. The measurements indicate that the proposed BGR can achieve a temperature coefficient as low as 1.01 ppm/°C over the temperature range of 160°C (−40 to 120°C).

A 2-tap pre-emphasis SST transmitter with skin effect loss equalisation is presented. The transmitter features the ability of compensating the low frequency loss caused by the channel skin effect. The transmitter equalisation coefficient setting and driver impedance adjustment are mutually decoupled. Fabricated in 65 nm CMOS technology, the transmitter can operate up to 20 Gbit/s while consuming 39 mW of power under a 1 V power supply. The skin effect loss equalisation is verified with a 24-inch FR4 channel under an 8 and 9 Gbit/s data rate.

A comparison of phase variation and insertion loss between a conventional digital attenuator and a proposed capacitive-phase-compensation digital attenuator design based on an analysis of phase variation is presented. In a standard CMOS technology, the proposed attenuator is analytically demonstrated to have less phase variation than the conventional digital attenuator due to the capacitive-phase-compensation network. The proposed switch can maintain <1 dB insertion loss while achieving an 8–18° improvement of phase variation than the conventional designs over 10–20 GHz, which support the theoretical analysis.

An energy-efficient 64 bit static sparse-tree adder using a multiplexer (MUX)-less bypassing scheme is proposed for mobile central processing units. Conventionally, bypassing schemes have been used to eliminate unnecessary switching of circuits but have incorporated a large delay overhead due to their output MUX, which reduces the energy efficiency of circuits in terms of power-delay product (PDP). A novel static sparse-tree adder is presented based on a proposed MUX-less bypassing scheme to reduce the delay associated with the conventional bypassing scheme, thereby improving the energy efficiency, i.e. the PDP. Simulation results show a 30% reduction in PDP compared to the conventional bypassing adder and a 13% reduction from the state-of-the-art technique.

The measurement results and a practical application of vertical inductors (VIs) are presented. VIs are on-chip inductors designed to save chip area. In VI the spiral plane is oriented perpendicularly with respect to the substrate. An inductance of 318 pH with a self-resonance frequency of 66.5 GHz is implemented in a 130 nm bipolar complementary metal–oxide semiconductor (BiCMOS) technology and results are compared with three-dimensional (3D) planar EM (ground–signal–ground) simulations. The inductance per unit area is 424 nH/mm², more than 1.7 times higher than other simulated spiral horizontal inductors. Moreover, VIs have been applied in the design of a broadband amplifier leading to a 25% bandwidth enhancement.

Voxelisation methods are extensively employed for efficiently processing large point clouds. However, it is possible to lose geometric information and extract inaccurate features through these voxelisation methods. A novel, flexibly shaped ‘supervoxel’ algorithm, called boundary-enhanced supervoxel segmentation, for sparse and complex outdoor light detection and ranging (LiDAR) data is proposed. The algorithm consists of two key components: (i) detecting boundaries by analysing consecutive points and (ii) clustering the points by first excluding the boundary points. The generated supervoxels include spatial and geometric properties and maintain the shape of the object's boundary. The proposed algorithm is tested using sparse LiDAR data obtained from outdoor urban environments.

Representation-based classification (SRC) is a face recognition breakthrough of recent years, but the dimensionality reduction for SRC has not been well addressed. The reason why existing dimensionality reduction methods are not effective for SRC is revealed for the first time. Based on analysis of the classification mechanism of SRC, the novel dimensionality reduction method for SRC is proposed, i.e. class coding error discriminant analysis (CCEDA), which simultaneously maximises the inner-class coding error and minimises the intra-class coding error. Extensive experiments show that the CCEDA feature achieves a better performance than the other features when using SRC as the classifier.

A novel face recognition approach by the use of a bundle of example images is presented, in which a combination of useful examples for face recognition is selected, and multiple augmented example images are made using those images. For this purpose, first, example images are divided into multiple groups using unsupervised methods such as the K-means method, and then clustering them into several groups according to their image variations. In each group, the most similar example images to an input image are found and they independently make an example average image from both the retrieved examples and their rank orders. When comparing similarities between two images, now local distance information can be utilised, which are calculated via corresponding average example images, as well as global distance information, the traditional direct comparison. The proposed approach is compared with the traditional direct comparison method on a multi-PIE database, and the generalisation ability of the proposed method across different features, such as local binary pattern, histogram of oriented gradients and Gabor features, is demonstrated.

An image segmentation approach, which is based on heat diffusion and graph cuts optimisation, is proposed. The prior segmentation result is obtained by temperature maximisation on the heat diffusion system. In the random walk-based label-assigning process, due to lack of spatial dependencies of neighbouring pixels, the segmentation may deteriorate notably when pixels from disconnected regions of an image show similar features. To overcome this problem, a multilayer graph-based model is presented and image segmentation is considered as an energy minimisation problem. The parameters in the model are learned from the results of temperature maximisation on the heat diffusion system. It is shown that the presented variational model can be discretely optimised by the graph cuts method efficiently. Therefore, the spatial dependences of the neighbouring pixels can be integrated to obtain better segmentation results. A number of comparison experiments demonstrate the superiority of the proposed method.

A driving waveform for electrophoretic displays with high image quality is proposed. The proposed driving waveform shows good performance compared to the conventional waveforms. By adding two voltage levels, the number of grey levels expanded from 8 to 165. In addition, the average response time improved from 90 to 61 ms.

Subspace clustering is very useful in many fields, such as computer vision and machine learning. However, most of the clustering methods cannot deal with out-of-sample data directly. For each new sample, these methods need to relearn the representations of all (new and original) data for clustering. This is unrealistic in many practical applications. A new online clustering method to cluster out-of-sample data in terms of the meaningful energy scores of data is proposed. By interpreting low-rank representation (LRR) as a dynamical system, a computation method for energy scores of data has been developed. The scores can be calculated by integration, independent of the LRR learning procedure. Then, a linear classifier is used to cluster out-of-sample data using their energy scores. Experimental results demonstrate the effectiveness and efficiency of the method.

For the feature extraction of red–blue–green (RGB) colour images, researchers usually deal with R, G and B channels separately to obtain three feature vectors, and then combine them together to obtain a long real feature vector. This approach does not exploit the relationships between the three channels of the colour images. Recently, attention has been paid to quaternion features, which take the relationships between channels into consideration and seem to be more suitable for representing colour images. However, there are only a few quaternion classifiers for dealing with quaternion features. To meet this requirement, a new quaternion classifier, namely, the quaternion softmax classifier is proposed, which is an extended version of the conventional softmax classifier generally defined in the complex (or real) domain. The proposed quaternion softmax classifier is applied to two of the most common quaternion features, that is, the quaternion principal components analysis feature and the colour image pixel feature. The experimental results show that the proposed method performs better than the quaternion back propagation neural network in terms of accuracy and convergence rate.

An efficient and effective algorithm which online exploits informative features for visual tracking is presented. First, a high-dimensional multi-scale spatio-colour image feature vector is developed, which takes into account both appearance and spatial layout information; secondly, this feature vector is randomly projected onto a low-dimensional feature space, where its projections preserve intrinsic information of the high-dimensional feature vector but effectively avoid the curse of dimensionality; and finally, an online feature selection technique to design an adaptive appearance model is proposed, which explores the most informative features from the projections via maximising entropy energy. Experiments on extensive challenging sequences demonstrate the superiority of the proposed method over some state-of-the-art algorithms.

A scale adaptive visual tracking algorithm based on the latent support vector machine (SVM) is proposed. The location of the object to be tracked is predicted by scanning all possible candidate locations and the scale is treated as a latent variable. With the predicted location, the latent SVM is optimised by a coordinate descent approach that optimises the latent variable and SVM parameters in an iterative manner. The separation of location and scale searching makes the tracker less likely to drift. Experimental results on test video sequences demonstrate that the proposed approach shows better accuracy than several state-of-the-art visual tracking algorithms.

Intrusion detection in real-time systems is a problem without a profound solution. In supervisory control and data acquisition (SCADA) systems the absence of a defence mechanism that can cope with different types of intrusions is of great importance. False positive alarms or mistakes regarding the origin of the intrusion mean severe costs for the system. An integrated one-class support vector machine (OCSVM) mechanism that is distributed in a SCADA network is presented, as a part of an intrusion detection system, providing accurate information about the origin and the time of an intrusion. The module reads the network traffic, splits traffic according to the source of the packets and creates a cluster of OCSVM models. These trained models run in parallel and can accurately and fast recognise different types of attacks.

The generation of intrinsic LLR messages in non-binary (NB) coded schemes associated with quadrature amplitude modulation (QAM) is considered. It is shown that the intrinsic LLR message generation corresponds to the same kind of computation as the one performed at the elementary check nodes in extended min-sum NB-LDPC decoders, i.e. finding a given number of minimum values in a structured set. It is proposed to use the bubble-check algorithm for the LLR calculation to benefit from two advantages: low-complexity hardware architecture and sharing the same hardware at the demapping and the decoding steps.

The problem of symbol misidentification (SMI) for data-dependent superimposed training (DDST) is considered. The constraint conditions on the discrete Fourier transform matrix are derived and constellation rotation (CR) at the transmitter to avoid the SMI is proposed. Simulation results show that the DDST with CR can eliminate the symbol error floor and yield better detection performance than the original one.

The exact pairwise error probability (PEP) of quasi-orthogonal space–time block codes (QSTBCs) employing the power scaling technique in quasi-static Rayleigh fading channels is derived. By utilising the derived PEP formula, the tight union upper bound on the bit error rate (BER) for the QSTBC with power scaling (PS-QSTBC) is evaluated and compared with the Monte Carlo simulation results to verify the analytical results that full diversity rate-1 transmission can be achieved. Furthermore, considering the closed-loop system, the optimal value of the power scaling factor is also investigated in terms of minimising the union bound on the BER.

A fast identification algorithm for Wiener-Hammerstein systems is proposed. The computational cost of separating the front and the back linear time-invariant (LTI) block dynamics is significantly improved by using discrete optimisation. The discrete optimisation is implemented as a genetic algorithm. Numerical results confirm the efficiency and accuracy of the proposed approach.

The performance of an energy detector operating over κ–μ shadowed fading is investigated. A novel closed-form expression for the average detection probability using the well-known probability density function-based method is derived. The unified framework, in terms of Fox's H-function, shows that the obtained results can be used for arbitrary values of fading parameters. On the basis of this derived analytical formula, the impact of key fading parameters on the performance of the energy detector is discussed. Furthermore, the accuracy of the analytical expressions is substantiated by Monte Carlo simulation results.

In projector-based applications, if the laser spot of a laser pointer can be quickly and accurately tracked, the laser pointer can act as an alternative to a computer mouse, and thus a user can more conveniently interact with the computer from a distance. Presented is an efficient calibration algorithm, which is immune to both the geometry of projection surfaces and the distortion of optical lens, for laser pointer tracking. Experiments verify the effectiveness of the proposed algorithm.

An investigation into new diplexer topologies, based on all-resonator structures with cross-couplings between common resonators (shared by both channels) and branch resonators is reported. This all-resonator structure eliminates the need for separate frequency distribution networks and uses resonators to achieve this functionality. For diplexers based on such all-resonator structures, cross-couplings can be added between the common resonators and branch resonators to achieve some desired specification (e.g. improved isolation). Two diplexer topologies with such cross-couplings are presented. The first topology is implemented at the X-band using waveguide technology. Excellent measurement results verified the proposed topology as well as the design procedure.

A compact three stub-loaded open-loop resonator-based tri-band bandpass filter with better out-of-band suppression is presented. The position and length of the three stubs loaded to an open-loop resonator determine the position of the three passbands. Mathematical analysis has also been performed to calculate the resonance condition of the proposed resonator. To demonstrate the concept, two second-order filters are fabricated having three passbands centred at f_o = 1.93 GHz, f ₁ = 3.6 GHz, f ₂ = 4.89 GHz and f_o = 1.59 GHz, f ₁ = 3.12 GHz, f ₂ = 4.02 GHz, respectively. The fabricated filters show better 3 dB bandwidth as well. The simulation and measured results are in good agreement.

A terahertz (THz) micromachined waveguide bandpass filter based on metallic two-dimensional (2D) electromagnetic crystal (EMXT) with a square lattice is proposed. The filter structure with three dual-mode circular cavities and two rectangular waveguides is obtained by removing and optimising the EMXT unit cells in a 15 × 20 cells array. This filter, fed by WR2.2 waveguides, is fabricated with the silicon microelectromechanical systems technique and measured using a vector network analyser (VNA) with an extender. The measured fractional bandwidth is 7.6% with a 426.5 GHz centre frequency. The best insertion loss and return loss are 2 and 24 dB, respectively. These results agree well with the simulations. The proposed THz EMXT filter indicates a novel approach to obtain THz waveguide passive devices and has application potential in THz communication and detection.

A lumped quadrature coupler based on a ferromagnetic core is proposed for high-frequency (HF) application. The four terminals of the transformer are connected to the four ports, whereas two inductors and two capacitors are connected between two adjacent ports. The bandwidth is only limited by magnitude imbalance. A 13.56 MHz prototype is designed and tested. A 69% bandwidth is achieved with ±3 dB magnitude imbalance. Within the frequency band, the measured return loss is greater than 31.57 dB, port-to-port isolation is greater than 33.21 dB and the phase imbalance is <2°.

A novel sub-harmonic upconversion mixer using Chartered 0.18 μm radio-frequency (RF) CMOS technology for 2.4 GHz wireless application is introduced. Using the sub-harmonic technique, the required local oscillator (LO) frequency can be reduced a half, which simplifies the design of the LO. Moreover, the sub-harmonic technique can effectively reduce the influence of LO leakage. The measured results demonstrate that the novel upconversion mixer can provide about 14.4 dB conversion gain, whereas the power consumption is only 1.65 mW with a 1 V supply voltage, and the input-referred third-order intercept point (IIP3) of the mixer is about 4 dBm. The chip area including test pads is only 0.7 × 0.72 mm².

A wideband single-layer absorber at X-band adopting trumpet structures is presented. One unit of the proposed absorber is composed of a trumpet-shaped resonator loading four chip resistors, a metallic backplane and an FR-4 (ε _r = 4.4-j0.02) substrate between them. The absorption is 99% at 10.9 GHz. The full width at half maximum is 95% at 12.7 GHz (6.7–18.7 GHz). The 90% absorption bandwidth is 65% at 12.75 GHz (8.6–16.9 GHz). The size of the unit cell is 5.6 × 5.6 × 2.4 mm³. The proposed absorber is almost insensitive to polarisations of incident waves due to the symmetry of the structure. The results of EM simulation and measurements are in good agreement.

A novel and unsplit-field implementation of the stretched coordinate perfectly matched layer (SC-PML) is proposed for truncating the one-dimensional (1D) frequency dispersive first-order Debye medium. These formulations are based on the unconditional stable Crank-Nicolson scheme and the Z-transform method. Numerical results validate the effectiveness and performance of this new PML.

The hysteresis in a nanoelectromechanical resonator is a common nonlinear characteristic that can be induced with excessive driving power. In this reported work, the hysteresis was controlled with the resonance beam motion amplitude changed by additional DC voltage which leads to the Joule heating effect. The mechanical properties of the beam which can be changed by the Joule heating effect appeared to be closely related to the hysteresis.

A double-gate charge-plasma-based super-steep negative capacitance junctionless tunnel field effect transistor (NC-JLTFET) using a ferroelectric gate stack is proposed. Structurally, the NC-JLTFET consists of a heavily doped n-type silicon (Si) channel with two distinctive gates (control gate and fixed source gate). The fixed source gate accounts for the charge-plasma (hole plasma) formation which results in surrogate p-type doping by using work-function engineering. It induces a uniform p-region on the source side on the n-type doped Si film having a thickness less than the Debye length (L _D). The key attribute of the NC-JLTFET is the ferroelectric gate stack which is employed as a control gate resulting in NC behaviour due to positive feedback among the electric dipoles in the ferroelectric material. The NC-JLTFET endeavours to achieve a super-steep sub-threshold slope, a paramount boost in drive current and a substantial enhancement in peak transconductance (g _m) than the JLTFET. Meanwhile, it embraces the inherent advantages of the charge-plasma junctionless structure. Thus, it avails itself of a simple fabrication process flow and high immunity against process variations and random dopant fluctuations.

A scanning frequency of 350 kHz is demonstrated with a full deflection angle of 95.8 mrad using a KTa_1−x Nb _x O₃ (KTN) deflector. To suppress degradation of the deflection angle, the KTN chip is directly glued to metal block electrodes with conductive paste, and the temperature of the deflector is controlled so that the capacitance of the KTN chip remains constant.

Multiplexed gallium–arsenide (GaAs)-based terahertz (THz) emitters fabricated with periodic double-metal structures are demonstrated and the effect of different metals on the THz output is investigated. THz emission originates from the lateral photo-Dember effect and from the different Schottky barrier heights of the chosen metal pair. The metal combinations used were Au–Al, Au–Pb and Cu–Cr. The emitters were characterised according to temperature and the highest peak-to-peak THz emission was achieved with the Cu–Cr metal pairing at 150 K.

A novel narrow-band tunable transmission filter based on a sub-wavelength metallic grating incorporating polymer-dispersed liquid crystal (PDLC) with a tunable range of 47.2 nm is demonstrated. The transmission peak wavelength can be tuned from 1524.9 to 1572.1 nm when the refractive index of the PDLC layer is varied by different applying voltages. More than 95% transmission efficiency can be achieved at the resonant wavelength, while the −3 dB bandwidth is about 1.6 nm. In addition, the resonant wavelength of the filter varies in a linear fashion with regard to the refractive index of the PDLC layer. Hence, a desired resonant wavelength can be selected easily by controlling the applied voltage.

Based on the optimised high-voltage network design and the resistance matching technique, a wafer-level light emitting diode (WL-LED) with a light output power of 157 W has been successfully demonstrated for the first time. The external quantum efficiency of the fabricated WL-LED was measured to be 24% at a driving current of 4 A.

A flicker-free transformerless light emitting diode (LED) driving circuit based on a quadratic buck power factor correction (PFC) converter is proposed. It combines a cascaded input buck PFC converter and an output buck dc–dc converter with a single switch. The proposed LED driving circuit can not only achieve high PFC to comply with the EN61000-3-2 standard and low output voltage without a high step-down transformer, but also reduces second-order line frequency current ripple flowing through the LED to eliminate light flicker of the LED. Experimental results of a 7 W prototype are presented to verify the analysis results of the proposed converter.

A simple empirical formula is proposed to calculate the self-resonant frequency of Archimedean spiral coils made of circular wire. A spiral coil can be considered as a deformed straight wire with some specific structural parameters. Each structural parameter is assumed to affect the self-resonant frequency. Two spiral coils, one that resonates near 13.56 MHz and one that resonates near 6.78 MHz, were fabricated to test the proposed formula. The average per cent error between the measured result and the calculated result was 2.66%. The proposed formula will be helpful in the design of Archimedean spiral coils for wireless power transfer.

The design of a frequency-modulated continuous wave (FMCW) radar altimeter with wide altitude range and low measurement errors is presented. Wide altitude range is achieved by employing the optical delay in the transmitting path to reduce the dynamic range of the measuring altitude. Transmitting power and receiver gain are also controlled to enable the dynamic range of the received power to be reduced. In addition, low measurement errors are obtained by improving the sweep linearity using the direct digital synthesiser (DDS) and minimising the phase noise employing the reference clock (Ref_CLK) as the offset frequency of the phase-locked loop (PLL).

The curve linear–linear (CLL) basis functions are proposed to improve the accuracy of both the magnetic-field integral equation (MFIE) and the combined-field integral equation (CFIE) for electromagnetic scattering from PEC objects. The CLL basis functions are derived first and then the multilevel fast multipole algorithm is applied to accelerate the matrix–vector multiplication. The accuracy of the CLL basis functions for both the MFIE and the CFIE are compared with both the curve RWG basis functions and the linear–linear basis functions. Numerical results demonstrate the efficiency of the proposed method.

A convenient formula for the estimation of clutter degrees of freedom is derived in wideband multiple-input and multiple-output (MIMO) radar with frequency diversity waveforms. The estimation of the clutter rank of MIMO radar is extended to the mode of the wideband with the help of the time-bandwidth product lemma; this could be valuable for designing an optimal technique in the application of parameter estimation, target recognition etc. The proposed rule is well verified by simulation results.

A self-terminating gate recess wet etching technique with thermal oxidation of the AlGaN/GaN layer followed by etching in potassium hydroxide (KOH) solution was recently proposed by the present authors for normally-off AlGaN/GaN metal–oxide semiconductor field effect transistors (MOSFETs). In this present reported work, the oxidation process inside the AlGaN/GaN heterostructure involved in this technique was analysed using several material characterisation methods. The measurement results show that the concentration and depth of the O element distribution increase with increased thermal oxidation temperature. It is worth noting that after 650°C oxidation almost no O element could be found in the GaN layer and the O element mainly locates in the AlGaN layer with an obvious correlation between the distribution of Al and O elements, where the Al(Ga)-oxide was detected by X-ray photoelectron spectroscopy, which could be etched by 70°C KOH. Thus, self-terminating wet etching on the AlGaN/GaN material is achieved.

A novel vertical field plate (VFP) structure with low specific on-resistance (R _on,sp) is proposed. The VFP is inserted in the field oxide of the drift region with heavily doped N pillar parallel to the trench oxide layer (TOL), which depletes fully the drift region to decrease R _on,sp effectively and enhances the bulk field (ENBULF). The VFP optimises the bulk electric field to increase the breakdown voltage (BV). The BV of VFP is 668 V with an R _on,sp of 44.7 mΩ cm², which is much lower than the silicon limit.

In a wireless communication system, the performance of the system is degraded due to the attenuation of the signal by path loss when a user is distant from the base station. For improvement of the system in this case, a new scheme is proposed. In conventional schemes, the source can transmit to the destination belonging to own cell coverage. In the proposed scheme, the destination can receive the additional signal from another base station through multiple relays. Therefore, the proposed scheme provides higher diversity gain and better performance than the conventional scheme.

A feature-based modulation recognition algorithm is proposed, to identify practical multilevel amplitude phase shift keying (MAPSK) signals using template matching, which suffers from various unknown effects, including carrier frequency, time delay and phase shift. The probability distribution functions (PDFs) of the amplitudes for each modulation constellation are estimated from the corresponding analytic signal of a received intermediate frequency signal. Then, a template matching algorithm based on Kullback-Leibler divergence is used to evaluate the similarity between the PDF of the amplitude for a received signal and the PDF templates, which has been stored in advance. The simulation results show that the proposed modulation recognition algorithm performs robustly and precisely even in an extremely low signal-to-noise ratio. The performance of the proposed algorithm is not very sensitive to the number of used symbols or timing error, which means that the performance will not be seriously degraded with a less number of symbols or with a larger timing error for symbol synchronisation.