A structure and computationally efficient design optimisation procedure of a novel miniaturised UWB slot antenna has been proposed. Upon geometry parameter adjustment, the structure exhibits a very small footprint of just 161 mm². This has been achieved by introducing sufficient number of degrees of freedom, including the parameterised multi-section slot and a multi-section load of the feed line as well as efficient numerical optimisation of all relevant geometry dimensions of the structure. The design process has been formulated as explicit size reduction problem. The electrical performance parameters (here, a maximum in-band reflection) are controlled by suitably defined constraints. The final design has been verified through full-wave electromagnetic simulation as well as physical measurements of the fabricated antenna prototype.

A simple single-feed resonant cavity antenna having a flat, low-profile all-dielectric partially reflecting surface (PRS) is described, which has a wide 3 dB bandwidth covering the entire Ku-band. The PRS is composed of three dielectric sections, forming a planar slab with the dielectric constant of the slab decreasing from the centre to the edge in three steps. In addition to the directivity enhancement (up to 10 dB over the feed antenna), it is shown that such single-layer PRSs remarkably improve the bandwidth of RCAs while reducing the profile of the antenna significantly. The total height of the antenna prototype is only 0.65λ ₀ at the lowest operating frequency. Experimental results are included to validate the proposed concept.

To multiplex multiple orbital angular momentum (OAM) beams, a compact half-mode substrate integrated waveguide leaky-wave antenna (LWA) for generating multiple high-order OAM modes is presented. The LWA is based on a cavity resonator with two feeds of ±90° phase difference. The principle and mode analysis of the resonator are discussed, and the generation of high-order OAM beams is proved by the measurements.

A method for detecting shockable rhythm was proposed. A single feature was extracted using the discrete wavelet transform from a single-channel electrocardiogram. The extracted feature was applied to the support vector machine to determine whether the signal was a shockable or non-shockable rhythm. To evaluate the performance, five-fold cross validation was performed with different window lengths (WLs). As a result, a sensitivity of 97.8% and a specificity of 98.0% with 3 s WL were obtained. The results represent the proposed method shows high detection performance even with short analysis window and only a single feature.

A new empirical mode decomposition (EMD)-based electrocardiograph (ECG) compressor is presented. Utilising the sifting process in EMD, the spline fitting function of the successive mean extrema together with the first intrinsic mode functions can perfectly reconstruct the ECG signal, the proposed method using the mean extrema for the compression. With optimal selection of the mean extrema and dead zone quantisation, and the application of run length coding and Huffman coding for further compression, the experimental results regarding several records of the MIT-BIH arrhythmia database shows that the proposed methods have a better compression results than recently developed ECG compressors.

A high energy-efficiency capacitor switching scheme for successive approximation register (SAR) analogue-to-digital converters (ADCs) is presented. The switching method, verified on a 10-bit SAR scheme that uses bottom plate sampling, achieves an average switching energy and area reduction of 99.32 and 96.5%, respectively, with respect to the conventional solution.

An inverted bit-line sense amplifier (BLSA) equipped with offset compensation capability for low-power DRAM applications is proposed. The sequential operation of the inverted BLSA allows us to eliminate the edge dummy array in an open bit-line structure resulting in 1.7% less total chip area despite of 10% area penalty of the proposed BLSA occupied by extra switches. For 8-Gb DRAM in 20-nm class technology, the read failure induced by Vth variability is completely removed due to the offset cancellation. The proposed BLSA maintains the gradual increase of the sensing delay when decreasing the power supply down to 0.6 V, while intrinsic read fail prevails below 0.9 V with the conventional one.

A fractional-N multiplying delay-locked loop (MDLL) with delay-locked loop (DLL)-based injection clock generation is presented. By exploiting multiphase output of DLL which delay is locked to the period of output frequency, the proposed architecture performs a fractional clock multiplication with MDLL, while eliminating deterministic jitter from fractional divider. The proposed MDLL is designed in a 0.18 μm CMOS process and achieves 31.25 kHz frequency resolution with 1 MHz reference frequency. It occupies an active area of 0.055 mm², and consumes 45 μW for 10 MHz frequency generation, showing energy efficiency figure-of-merit (FoM) of 4.5 μW/MHz.

The solution of a problem of circuit optimisation for minimum possible CPU time is obtained on the basis of the maximum principle of Pontryagin. It is shown that the effect of acceleration of the optimisation process that was studied earlier coincides with the solution obtained on the basis of the maximum principle. This is a theoretical justification of the effect of acceleration. On the other hand, the maximum principle can be a basis for development of optimisation algorithm for electronic circuits with the minimum expenses of CPU time.

A multi-step incremental ADC (IADC) with extended binary counting is proposed. It achieves high accuracy by splitting one conversion cycle into two serial steps. During the first step, the ADC works as a first-order IADC (IADC1). The second step reuses the single integrator and extends the accuracy to 18 bits by a binary counting technique. For the same accuracy, the conversion cycle is shortened by a factor of more than 2⁸ as compared with the single-step IADC.

An energy-optimal coverage path planning algorithm is proposed for environment survey. First, 3D terrains are modelled by tensor product Bezier surfaces for a mesh representation. Then, a power estimator is derived to calculate the consumed energy for a piecewise spatial path. Based on the digital surface model, an energy consumption map is constructed for the whole area by means of a weighted directed graph. Finally, an energy-optimal path can be achieved through traversing the map by a genetic algorithm. Numerical experiments demonstrate the effectiveness and efficiency of the proposed algorithm.

A directional and omnidirectional edge enhancement method is proposed based on radial Hilbert transform (RHLT) of directional Gabor filter (DGF). A new transfer function is introduced to develop a new RHLT that can be rotated to enhance edges along the specified direction. Furthermore, to increase the edge response at the specified direction and suppress noise, DGF with the centre frequency being at the direct current component, is designed to be radial Hilbert transformed.

A new non-parametric inferring procedure of facial feature localisation which learns regression in latent variable space is introduced. The proposed method uses regression between latent feature and motion spaces spanned by a set of bases of nonlinear feature space and motion vector space which is trained from a large sample set in codebook. Compared with the previous method of using codebook or eigen-codebook, the proposed method achieved both a significant reduction in the memory consumption without loss of accuracy and a very low computational complexity enough to run in mobile devices.

A low-noise analogue frontend readout circuit is presented for VOx microbolometer infrared imager without thermo-electrical cooler. A modified differential mirror structure with substrate temperature and self-heating effect compensation ability is used, targeting below 40 mK noise equivalent temperature difference (NETD) with 17 μm pixel pitch. Noise characteristics are investigated to give an optimal trade-off approach to reach lower NETD at same power consumption compared with traditional readout structures.

Two different subgroups C _n and L _n of symmetric group S _n are first designed and then to construct a family of permutation group codes P _n with code length n, minimum Hamming distance n − 1, cardinalityn(n − 1) and error-correcting capability n − 2 is composited.

An adaptive and recursive approach to the network traffic modelling taking into account multifractal characteristics of the network traffic flows is proposed. To this end, equations for some statistics of the scaling coefficients and for the model multipliers in order to develop a new algorithm that enables the adaptive estimation of parameters of a multiplicative cascade in the wavelet domain are presented. Simulations and comparisons with other three important models in the literature were carried out in order to verify the performance of the proposed approach.

Holographic data storage (HDS) provides the high storage density and fast data-access rate because of the page-oriented data processing, and by the same reason, the performance of retrieving original user data is mainly affected by 2D intersymbol interference (ISI). An 8/10 2D modulation code for mitigating a 2D ISI in HDS systems is proposed. The simulation results show that the proposed code provides a gain of approximately 6 dB at BER of 10⁻⁶ compared to the raw signal without encoding. Especially, in comparison to a recently suggested 6/8 modulation code, the proposed scheme not only offers a higher code rate, but it also exposes a boost of the system performance under misalignment conditions.

A new class of bias-compensated affine-projection-like (APL) algorithms is proposed, in which a bias-compensation vector is derived to eliminate the bias caused by the noisy input. In addition, a new estimation method for the input noise variance is proposed which does not require the input–output noise variance ratio in advance. Simulations in a system identification context show that the proposed algorithms achieve significant improvements in steady-state misalignment as compared with the conventional APL algorithms.

A novel design strategy for adaptive impedance matching network which consists of two tunable components is presented. The input resistance is controlled by the first tunable component and independent of the second tunable component. Additionally, the proposed strategies ensure that the input resistance is proportional to the first tunable component and input reactance is proportional to the second tunable component. Therefore, the binary search algorithm is employed to calibrate the tunable matching network. In contrast to the single-step method, the convergence speed is improved from O(N) to O(log N).

Proposed is a blind recognition method for Bose–Chaudhuri–Hocquenghem (BCH) codes in the faster-than-Nyquist (FTN) signalling system which is based on the Galois field Fourier transform of the code polynomial. Considering the signal-to-noise ratio estimation in coloured noise, the proposed method can maximise the correct recognition probability of the roots of the generator polynomial by an appropriate decision criterion. Moreover, an optimal threshold of the decision criterion is deduced. Simulation results demonstrate that the proposed method can recognise the BCH codes efficiently in the FTN signalling system.

Direct computation of pth roots in extended finite fields of characteristic p ≥ 2 is introduced, wherein the reduction polynomial is irreducible and can be even random. Proposed method works in any case of p ≥ 2 and finite field extension. This method is the most efficient, it is even more efficient than the method, which is widely used, based on inversion of squaring matrix utilised in the case of p = 2. This method is more efficient regarding the computation and storing of the matrix as well as the computation of the roots.

Two constructions of a low-complexity near-optimal detection method of dc²-balanced codes are presented. The methods presented are improvements on Slepian's algorithm for optimal detection of permutation codes.

A novel threshold-based scheduling scheme for multiuser downlink wiretap networks is addressed, which consist of a base station (BS), multiple legitimate users and multiple eavesdroppers. Particularly, the active scenario is concentrated, where BS has access to the channel state information for the eavesdropper's channel. In doing so, new exact and asymptotic closed-form expressions for the ergodic secrecy capacity are derived for multiuser downlink networks using switch-and-examine combining with post-selection scheduling at legitimate users and maximum ratio combining at the eavesdroppers. Moreover, the high signal-to-noise ratio slope and power offset are also given to explicitly characterise the impact of main link and wiretap link on the ergodic secrecy capacity. Simulation results show the effect of key parameters such as the number of legitimate users and eavesdroppers, switching threshold on the secrecy performance of the considered wiretap network.

A novel dual-band bandpass filter (BPF) based on a single triple-mode ring resonator is proposed. The proposed BPF could provide three transmission poles for each passband to improve operating bandwidth. Meanwhile, seven transmission zeros are achieved at the stopband to improve the selectivity. The proposed dual-band BPF was designed, fabricated and tested. The measured results are in good agreement with the simulated values.

A novel simple broadband via-free microstrip composite right and left hand transmission line (CRLH-TL) is proposed by etching a straight-line slot in the conductor of the microstrip line and an eight-edge-shaped slot on the ground below the straight-line slot. The CRLH-TL has the advantages of simple structure, smaller size, and wide left hand frequency band. The results show that the left hand frequency band of the CRLH microstrip transmission line is 4.91−6.09 GHz, and the left hand frequency bandwidth is 1180 MHz.

In this Letter, a method for the broadband measurement of complex permittivity using a quasi-optical technique is presented. After calibration of the quasi-optical system, the material under test is inserted into the system and S-parameters are measured from 220 to 330 GHz using a vector network analyser. An extraction method is then applied to calculate the complex permittivity. The permittivity measurements are validated by comparing with a well-established extraction method and the comparison shows good agreement.

A tri-band frequency selective surface (FSS) with close band spacing based on split ring slots for y-polarised incident waves is investigated. The close band response is obtained by reducing the strong mutual coupling between the closely packed concentric rings through splitting the ring slots and redistributing the location of the surface currents at resonances. To validate the proposed FSS, a Ka band FSS was designed, fabricated and tested. Experimental results showed closely spaced resonances at 27.3, 31.7 and 36.2 GHz. Frequency ratios of 1.16 for the second to the first resonance, and 1.14 for the third to the second one were obtained. Also, good stability is observed for oblique incidence.

On the basis of a single multimode resonator (MMR), a miniaturised microstrip sext-band bandpass filter (BPF) with high band-to-band isolations has been presented. The proposed MMR is an stepped impedance resonator (SIR) with one short-circuited stub loaded in the middle and two identical open-circuited stubs located at two sides. Transmission poles (TPs) and transmission zeros (TZs) can be excited when the electrical lengths of two open stubs increase. The TZs introduced by open stubs separate the TPs into multiple groups; therefore, the design of a sext-band BPF becomes available. Moreover, two input/output tapped branches and radical stub-loaded short-circuited lines are treated as perturbation structures in order to increase design freedoms for improving filter performance. Good agreement between simulated and measured results validates the design method.

A new compact balanced ultra-wideband (UWB) bandpass filter using half mode dumbbell defected ground structure (DGS) is presented. To realise the UWB response, the ‘T’ shape multiple-mode resonator with short-ended stub is employed. To achieve wideband common mode (CM) suppression and compact design simultaneously, the concept of half-mode DGS is introduced in the two stacked microstrip layers. Afterwards, a balanced UWB filter is designed and measured for verifying the feasibility of the presented design method. The measured results match well with the expected results which indicate the proposed filter exhibits better than −30 dB CM suppression within the DM passband covering from 3.1 to 10.6 GHz.

Hybrid metal insulator metal plasmonic waveguide has been investigated at 1550 nm wavelength using frequency-domain electromagnetic solver of computer simulation technology microwave studio. This waveguide overcomes the limitations of conventional plasmonic metal insulator metal waveguide and provides good balance between propagation length and confinement. Propagation length up to 2.5 mm and confinement factor more than 80% have been observed. Waveguide bends such as 90° and S-shaped, based on the presented waveguide have been examined; these bends exhibit more than 90% transmission efficiency in a confinement area of 40 × 200 nm². The application of these structures in the design of chip scale integrated circuitry is motivated.

A novel miniaturised wideband bandpass filter (BPF) based on harmonic suppressed dual transmission lines (DTLs) is introduced with enhanced upper stopband rejection and passaband selectivity. The transmission and harmonic suppression characteristics of series and parallel connected DTLs are analysed adopting even–odd mode method. To verify the proposed structure, a high selective wideband BPF centred at 1.86 GHz with a fractional bandwidth of 93.5% is designed and fabricated. The upper stopband is extended up to 8.8 GHz (4.7f ₀) with the suppression level of above 35 dB. Furthermore, a size reduction of 30% can be achieved in proposed BPF, compared with conventional wideband BPF based on quarter-wavelength transmission lines.

An optical two-tone signal with 60 GHz frequency spacing is experimentally demonstrated, by sinusoidally modulating two lightwaves with different modulation indices and RF phases. By superposing two modulation lightwaves with orthogonal polarisation, polarisation orthogonality is achieved between undesired low-order optical sidebands and the desired high-order optical sidebands together with the incident light-wave. This enables us to suppress the undesired sidebands using polarisers instead of optical filter. Such operation has been implemented using optical modulator possessing RF ports for external termination, which is nested in one mode of polarisation-maintaining Sagnac interferometer (PMSI). This implementation is equivalent to optical single-sideband modulator, but required nested Mach–Zehnder structure is just one. Owing to this reduction together with use of one mode in PMSI, long-term stable operation more than 6 h has been successfully obtained, with 29 dB suppression ratio of undesired first-order sidebands against desired third-order optical sidebands.

The transmission of multiple standard-compliant long-term evolution advance (LTE-A) bands together with a 4-pulse amplitude modulation (PAM) baseband signal over 50-m-long 1 mm core diameter polymethyl methacrylate graded-index plastic optical fibre is demonstrated. Transmission of eight LTE-A 64-QAM bands and a 1.2 Gbit/s 4-PAM baseband signal over the fibre was achieved resulting in an error vector magnitude <8% and pre-forward error correction (FEC) BER < 10⁻³, respectively.

A high DC gain self-cascode structure of operational transconductance amplifier (OTA) design with bandwidth enhancement is proposed. Based on the concept of self-cascode structure, which provides high output resistance and transconductance, the proposed OTA presents a compensation methodology to obtain better unity-gain bandwidth. The proposed OTA is compared with previous structures, with the same size of input/output transistors and compensation capacitance as well as load capacitance, and simulation results show a DC gain enhancement of >20% and 2.4 times unity-gain bandwidth at the same time, and the performances of the proposed OTA bring a better figure of merit (FOM).

As the scale of graphene-based non-volatile memory is reduced, the ratio of access resistance R _A to total channel resistance R _TOT is increased. To investigate the effect of the R _A on I–V characteristics, we fabricated devices with various access lengths L _A and self-aligned structure. Proposed structure using self-aligned gate minimises L _A, and thereby improves the drain current, ‘on/off’ current ratio I _ON/I _OFF and transfer characteristics. In proposed structure, ‘off’ current is increased from 0.16 to 0.28 mA because R _TOT was reduced; ‘on’ current increased from 0.35 to 0.72 mA, but I _ON/I _OFF increased from 2.18 to 2.57. Proposed structure also had larger memory window (8.5 V) than did conventional devices (6.7 V).

A 0.35 μm BiCMOS current generator suitable for optical quantum random number generators is presented. The time dependence of the current pulses is of the form T/(T − t). The design uses the bipolar junction transistor translinear loop with two inputs supplied by a DC and one input supplied by a sawtooth current. The relative error of the measured waveform at 20 MHz is <3% in 97.5% of the time interval in which the current pulses have a rising slope. The current pulses can be generated up to a frequency of 100 MHz.

A high-efficiency multi-string LED driver based on constant current bus with time-multiplexing control scheme is proposed and analysed. Each LED string is connected to the constant current bus, which is provided by the output of a front-end high-efficiency converter. The current of each LED string can be independently controlled by the turn on duty ratio of its time-multiplexing switch. A capacitor is connected to the current bus to provide a current flowing path for constant bus current when none of the LED string is connected to the current bus. It needs only one output filter capacitor, which makes circuit volume small and cost low. Finally, a 7.2 W two-string LED driver is provided to verify the studied results.

Phase-locking of two longitudinal oscillating modes of a dual-mode microchip laser that has 101 GHz longitudinal-mode spacing has been successfully demonstrated by using electronic negative-feedback control. Since the two longitudinal modes were oscillating simultaneously in the same laser cavity, the common-mode fluctuation cancels, resulted in the small required control bandwidth of ∼4 kHz. This narrow-bandwidth phase-locking technique is useful for generating high-spectral-purity millimetre-wave signals with a considerably simpler setup.

A new simple analytical and closed-form solution for the harmonic content injected in the five-level inverters at low switching control technique is proposed. The equations of the fundamental and the third harmonic are formulated and solved. The switching angles have been calculated to fix the fundamental voltage at a specified value and cancel the third-order harmonic. Two approaches have been proposed based on the modulation index. The proposed method can be feasible for a wide range of modulation index.

An energy harvesting (EH) and information processing network based on two-way multiplicative relay is proposed. In the proposed scheme, two nodes communicate with each other via an energy constrained multiplicative relay node. The relay node uses power splitting-based relaying protocol for EH and information processing. The numerical results show that the proposed scheme outperforms the legacy amplify-and-forward EH relaying technique.

Flying-capacitor (FC) three-level (TL) DC–DC topology is an attractive candidate for the switching-mode power supply due to its low-voltage stresses on semiconductors, small size of filter and common-ground feature. However, to avoid the semiconductors to be damaged from an initial high voltage, the FC must be precharged to half of the input voltage by a bulky precharger. An idea to realise the precharging function by adding only one switch and cooperating it with the existent circuits is proposed. In this way, the necessary soft-starting circuit and Resistor-Capacitor-Diode (RCD) snubbers for reliability purpose can be fully utilised to precharge the FC. Merits of the proposed idea include the lower cost, the more compact structure and the higher overall efficiency. The principles of the proposed function-integrated network are analysed, and a 1.2 kW FC-TL conversion system is also built to verify the analysis.

A simple electronic circuit for the characterisation of high-intensity discharge (HID) lamps in the frequency domain by obtaining the transfer function of the lamp impedance is presented. The circuit uses a low cost power amplifier to apply a sinewave to the current and the voltage in the lamp. Also, this circuit employs an ordinary igniter based on a high voltage transformer. This circuit is a good alternative when a ballast designer prefers to get a lamp model from experimental Bode plots, and when a low budget is available to build a test bench to characterise HID lamps. Simulations and experimental results obtained from an HPS lamp are shown.

A new method to regulate the output voltage of the secondary side of an inductive power transfer (IPT) system through a DC-voltage controlled variable capacitor (DCVC) is proposed. As the equivalent capacitance of the DCVC can be controlled by a DC voltage, when it is used as a parallel tuning capacitor at secondary side of an IPT system, the output voltage can be controlled by this DC voltage through the effect of tuning and detuning. Further, a simple proportional plus integral control (PI) controller can be employed to generate the controlling DC voltage conveniently to make the output voltage constant. As there is no active switch used in the DCVC, it has no switch driving problems such as a high up to 20 V voltage source which is of particular advantage for its application at the secondary side of an IPT system. The DCVC also has the potential to be used at much higher frequencies than traditional switch mode capacitors because of the smoothness of the adjustment of its equivalent capacitance through a DC voltage.

The stable normalised least-mean fourth (SNLMF) algorithm is developed in the passive bistatic radar (PBR) system to realise the interference cancellation. The larger power of the reference signal may lead to the poor convergence rate. By replacing the reference signal vector with the instantaneous echo signal, the concept of the clutter attenuation (CA) is introduced in the SNLMF algorithm, and the CA-SNLMF algorithm is proposed to provide a fast convergence. The real measured data collected from the PBR system demonstrate that the proposed CA-SNLMF algorithm has a faster convergence speed and better cancellation performance than the SNLMF algorithm.

A multiple input multiple output (MIMO) radar allows its antenna elements to transmit multiple signal waveforms. This waveform diversity offers enhanced flexibility in transmit beampattern synthesis and signal design. Two algorithms where the transmit is synthesised beampattern by optimising transmitted signal covariance matrix is proposed. In the first algorithm, a method to achieve a desired beampattern by reformulating the design problem as an unconstrained semi-definite programming problem is proposed. Second, a method for beampattern synthesis such that the cross correlation beampattern, which is the correlation between the signals at different target locations, is minimised and at the same time the sidelobe levels are constrained to be lower than a threshold value proposed. These design problems are modelled as convex optimisation problem and are solved using Matlab based toolbox CVX.

The optimised limit for polarimetric calibration of fully polarised synthetic aperture radar systems is derived by establishing an error model as a function of cross-talk, channel imbalance and system noise. Compared with noise equivalent sigma zero, the polarimetric error below the optimised limit is too small to affect the signal of cross-polarised channel. Thus, polarimetric calibration could be relaxed or even ignored in this case. With the backscatter model, optimised limits for cross-talk and channel imbalance at X, C and L-bands are presented. Moreover, when ignoring channel imbalance, the limit for cross-talk is given in a quantitative way. These results are very useful in practice, allowing significant reduction in calibration cost.

Imaging radar with a suboptimal aperture length usually has a poor azimuth resolution due to the limited aperture length and sparse non-uniform linear array configuration. To address this problem, an imaging algorithm combining pseudo-polar formatting and gridless sparse recovery method is proposed. After pseudo-polar formatting the echo signal, reweighted atomic norm minimisation is implemented for azimuth imaging to achieve performance with improved resolution and free of off-grid effect. Experiments on ground-based synthetic aperture radar verify the performance of the proposed algorithm.

Wideband radar cross-section (RCS) reduction of a microstrip patch array antenna is proposed and investigated. The design is based on replacing the solid metal ground of the antenna with a frequency selective surface (FSS) consisted of miniaturised hexagonal loop elements. The stop band of the FSS is coincident with the antenna working frequency band which is from 3.3 to 3.4 GHz. The FSS acts as the ground plane in the stop band and is transparent to the incident waves in the other frequencies. Therefore, preserving the antenna propagation characteristics, wideband RCS reduction is achieved from 4 to 16.5 GHz for vertical and horizontal polarisations. It is shown that using this FSS has a negligible effect on the impedance bandwidth, pattern and gain of the antenna. Also, there is a good agreement between the simulation and measurement results.

An accurate dead-time adjustment method for phase-shifted full bridge (PSFB) converter is presented. This method allows the dead-time of converters to be optimised to meet zero voltage switching condition. The optimum dead-time range can be predicted with high accuracy for any type of power semiconductor switch, including gallium nitride high-electron-mobility transistor, so that a practical design for a PSFB converter can be realised. The validity of the proposed dead-time optimisation method was verified by experimental results with a 600 W prototype.

A novel approach to suppress the ambipolar behaviour and enhance RF parameters is proposed for the first time. For this, the dielectric and gate material work function engineering is used to suppress the ambipolar behaviour individually. Further, the combination of gate dielectric and gate material work function engineering is used to suppress the ambipolar conduction in huge amount and to eliminate the hot carriers effects. Apart from these, the proposed work improves the ON-state current and RF figures of merit for symmetric devices.

A novel discrete Gabor transform (DGT) with a weighted linear combination of analysis windows is presented, which can utilise the optimal weights to change the shape and width of the combined analysis window so as to improve the time–frequency resolution of the Gabor spectrum. The optimal weights can be solved by minimising the objective function with the mixed ℓ₁ and ℓ₂ norms of the DGT coefficients, which leads to the maximum time–frequency concentration of energy on the time–frequency plane, or equivalently, the optimal time–frequency resolution. Experimental results show that the proposed DGT is highly effective in time–frequency analysis.

A two-path multiple-input multiple-output (MIMO) relay system is addressed, in which the inter-relay interference (IRI) between two relays leads to deterioration of system performance. Transmit and receive weight vectors are employed for suppressing the IRI. Maximising the ratio between the power of the desired signal to the IRI power at the relay reception and transmission and forming a criterion function for jointly calculating the optimal receive and transmit suppression vectors, respectively. Simulation results show that the proposed scheme can improve the system performance.

An adaptive modulation and coding (AMC) scheme based on a Moore state machine is proposed for the high-speed railway communication systems. By setting a buffer zone for the signal-to-noise ratio threshold value, the proposed scheme is able to reduce the frequent transition of the modulation and coding scheme. Simulation results show that the proposed AMC strategy has a more stable spectral efficiency than the traditional AMC method using a piecewise function.

Massive multiple-input multiple-output (MIMO) systems are being proposed for next generation broadband wireless systems. However, as the complexity of implementation grows with the number of antenna elements, their feasibility is a challenging task. Consider very low complexity, 1-bit digital-to-analogue converters (DACs) for the downlink of massive MIMO systems. The impact of the resultant severe non-linear distortion effects is analysed when a low-complexity maximum ratio transmission technique is employed for user separation. The non-linear distortion levels decrease with the number of transmit antennas allowing good performance, even with 1-bit DACs are shown.