Justification for the use of fragmented elements in the design of mosaic reflectarrays is provided whereby the far-field pattern performance is shown to be good in spite of the optical mosaic image, and this fact is confirmed for an additional test case of a synthesised image. Additionally, the use of sub-wavelength spaced elements, as opposed to the traditional half-wavelength spaced elements, is strongly motivated by sampling density advantages for image reproduction.

A reconfigurable ring antenna for switchable circular polarisation (CP) is presented. The proposed antenna consists of a square ring resonator with two gaps, two pin diodes and a square ring slot with a link on the ground plane. Depending on the location of the gap (link), the square ring (slot) resonator exhibits different directions of linear polarisation. By selecting a suitable gap/link position, two orthogonal resonant modes are excited, and good CP is achieved. The pin diodes are positioned in the middle of the gap. The state of the gaps can be adjusted by pin diodes' state. As a result, the proposed antenna can radiate with either right-hand CP or left-hand CP.

A multiple-input–multiple-output (MIMO) antenna system for mobile handsets is presented. The proposed MIMO system consists of four identical box-folded planar inverted-F antenna. A broad operating band (1.84–2.69 GHz) with overall MIMO system dimensions of 136 × 68.8 × 6 mm³ and an element size of only 10 × 10 × 5 mm³ can be achieved. Lower than −14.2 dB isolation between any two of the elements has been obtained at the whole operating band. The correlation coefficient, radiation pattern and antenna gain are also studied.

The design of an ‘orbital angular momentum (OAM) antenna’ based on an original phase shifter and power divider combination at 2.5 GHz is presented. A prototype is manufactured and characterised based on return loss, radiation and phase patterns. Both simulation and experimental results evidence the properties of waves carrying OAM. This OAM antenna is simple, compact and easy to realise, which makes it suitable for radio communications and radar applications.

A robust adaptive beamforming technique based on a reconstructed covariance matrix and array steering vector (ASV) estimation using low-complexity algorithms is proposed. The spatial spectrum estimator is used to reconstruct the interference-plus-noise covariance matrix and the ASV estimation process is used to calculate the correlation coefficients of the assumed ASV and the eigenvectors. Contrary to other robust adaptive beamforming methods, this approach has better performance yet does not involve any specific optimisation programme even if arbitrary ASV imperfections are considered. Simple implementation and significant signal-to-interference-plus-noise ratio enhancement support the practicability of the proposed method.

A wideband aperture-coupled patch antenna is presented. The proposed antenna consists of patches in different sizes and a ground plane, which is fed by a microstrip line through a slot cut on the plane ground. Owing to the patches of different size and the slot on the ground, three resonance modes are excited simultaneously for the radiation at the boresight. The proposed antenna yields a wide bandwidth of about 31% with a low profile of 0.07 λ ₀ ( λ ₀ is the free space wavelength at the centre operating frequency). The results are validated by both numerical simulation and experimental measurements.

A technique to remove baseline wander (BW) from electrocardiogram (ECG) signals based on the Hilbert vibration decomposition (HVD) technique is presented. It is proposed that the first component (highest energy component) obtained using the HVD of the ECG signal corresponds to the BW signal in it. The proposed technique is compared with the baseline removal method based on the empirical mode decomposition technique and mathematical morphology in terms of the correlation criterion and signal-to-noise ratio. The simulations were performed using artificial BWs of different amplitudes added in the ECG recordings and it is seen that the proposed technique for BW removal performs better in most of the cases of severe baseline wander distortions.

A novel three-dimensional (3D) electrical impedance mammography (EIM) system, named the combined EIM (CEIM) system, is introduced. It combines a ring electrode array with a planar electrode array that can operate independently or together. Three operational modes of the CEIM system are presented. Using a close-to-realistic digital breast phantom, the CEIM system is shown to provide enhanced detectability of a tumour deep within the tissue, as required for a large-volume breast. An implicit comparison with the authors' former MK4 system demonstrates the benefits of this integrated multimode mammography system.

In medical ultrasound imaging, the focused echo signal is decimated with a finite impulse response filter by a rational factor M/L, the frequency response of which dynamically changes over time. Presented is, an efficient structure for dynamic decimation by a programmable rational factor that use a fixed number of polyphase multiplier-accumulators (MACs) regardless of the values of M and L. The proposed structure uses the minimum number of MAC units operating at the input data rate to minimise both power consumption and hardware complexity.

A non-invasive piezoelectric-based sheet sensor is utilised for the deliberate measurement of heartbeat rate and blood pressure. It may consist of a single sheet to capture various temporal signals or it can be formed as an array of small piezoelectric sensors to capture the temporal and spatial cardiac signals over the chest to give an added spatial granularity on top of the localised temporal signal. The mechanical activity of the heart will cause the piezoelectric sheet to deform, and thus produce a voltage signal. To count the heartbeat rate, a mechanical model has been developed using the corresponding electric waveform; a compression ratio is defined due to the delay as a result of the physical coupling mechanism. Validation of the obtained results has been performed using conventional techniques and apparatus. The quantitative assessment shows low errors when compared with other conventional meters.

A three-stage amplifier employing embedded capacitor-multiplier compensation (ECMC) and active parallel compensation (APC) to enhance the area efficiency when driving nF-range capacitive loads (C _L) is presented. Unlike the conventional current-buffer Miller compensation, ECMC applied to the dominant compensation path saves substantial power and area, while securing a large gain–bandwidth product. The created left-half-plane zero also benefits the phase margin (PM). For the APC, unlike the traditional passive parallel compensation, this work benefits from the Miller effect to avoid the area-consuming resistor, and reduces the entailed capacitances without lowering the parasitic pole position. A multi-path G _m-boosting second stage enhances the effective transconductance and DC gain. With 0.0045 mm² of area and 15.8 µW of power, the 0.18 µm CMOS three-stage amplifier measures 1.13 MHz unity-gain frequency, 0.41 V/µs average slew rate and 56.2° PM at 1 nF C _L. Stable responses with >50° PM are attained for a 10 × range of C _L from 0.15 to 1.5 nF. The achieved figure-of-merit accounting for both die area and power compares favourably with the state of the art.

A new circuit technique is proposed to enhance the locking range of divide-by-3 injection-locked frequency dividers (ILFDs) implemented using a standard 0.18 μm CMOS process. The ILFD uses a transformer feedback to increase the second harmonic and enhances the conversion efficiency of injection metal–oxide semiconductor field-effect transistor mixers. The ILFD core consumes 10.8 mW power at the supply voltage of 0.9 V and with a circuit core current of 12 mA. At the incident power of 0 dBm, the measured locking range is 4.2 GHz (37.17%), from the incident frequency 9.2–13.4 GHz.

A new solution for a bulk-driven ultra-low-voltage operational transconductance amplifier (OTA) is described. While implemented in a standard 0.18 μm CMOS process, the circuit provides linear operation (total harmonic distortion <1%) for rail-to-rail input swing and supply voltages ranging from 0.25 to 0.5 V. The good linearity performance is Maintained over a large range of transconductances and biasing currents.

An energy-efficient capacitor-splitting digital-to-analogue converter (DAC) scheme with high-accuracy for successive approximation register (SAR) analogue-to-digital converters (ADCs) is presented. The proposed method uses a split-capacitive-array DAC structure and optimises the switching energy during conversion using energy-efficient ‘up’ transition. The proposed switching scheme achieves a 96.91% switching energy reduction and a 75% area reduction compared with the conventional method. In addition, the third reference voltage (V _cm) has no effect on the accuracy of the SAR ADC except the least significant bit, resulting in a good trade-off between the energy-efficiency and accuracy of the ADC.

An improved memristive Chua's circuit is proposed, from which some hidden attractors are found. This memristive circuit has three determined equilibrium points, including a zero saddle point and two non-zero saddle-foci. The two non-zero saddle-foci are stable for the specified circuit element parameters, leading to the generation of hidden attractors and coexisting hidden attractors. Additionally, the observed hidden attractors are not sensitive to the initial conditions, which can easily be captured in a physical circuit by adjusting the circuit element parameters to the desired values.

A new design scheme intended to improve the performance of true single-phase clocked (TSPC) dual modulus prescalers is presented. Two branches of TSPC D flip-flops are merged to reduce both power and device count. An HSPICE simulation of the proposed scheme demonstrates the highest power efficiency and best power–delay product among the referenced designs.

A high energy-efficiency mixed capacitor switching scheme is proposed for the successive approximation register (SAR) analogue-to-digital converter (ADC). The new switching scheme combines a new switching method, monotonic switching method and split-capacitor technique. The new switching method achieves no switching energy consumption until the fourth comparison, and the low-power monotonic switching method is utilised for the remaining comparisons. Moreover, fewer bit capacitors are needed at the same resolution because of the split-capacitor technique. The proposed method achieves a 99.4% reduction in switching energy when compared with the conventional SAR architecture.

A sliding-mode speed controller for output tracking of underactuated surface vessels is presented based on a new manifold definition. The new sliding-mode control algorithm relies on a conventional sliding manifold and its time derivative. The states/outputs are estimated with the extended Kalman filter as an observer. The new controller is compared with its integral-type sliding-mode counterpart in terms of performance in tracking the outputs where the estimator is driven by respective control inputs and observed states involved with arbitrarily correlated process and measurement errors.

A practical high curvature path planning algorithm that considers the velocity limit of a mobile robot in joint space is proposed. The existence of obstacles that constrict the smooth movement of mobile robots is inevitable. To avoid incoming collision, the robot is redirected to a new path with respect to the dimensions of the obstacle and a safety factor. The redirected path consists of geometric constraints such as high curvatures on its turning points which cause difficulty in controlling the mobile robot. Thus, a central velocity generation method that considers both velocity limits through a convolution operator and high curvatures at turning points is suggested. The central velocity is able to be within the configured velocity limit, otherwise actual velocity commands in joint space could result. Therefore, interpolation to obtain uniform sampling and velocity downscaling is performed to generate a feasible joint space trajectory. The result shows short travelling time compared with previous methods while satisfying velocity limits in the joint space.

Scanned laser pico projectors utilising laser light sources and microelectromechanical systems mirror devices introduce geometric distortion to the displayed images because of nonlinear mirror driving methods. A pre-processing algorithm is proposed that corrects the nonlinear geometric distortion of scanned laser pico projectors. The distortion is modelled as blockwise homography transforms. The geometry of the images being displayed is pre-distorted using the pixel indices obtained by the backward blockwise homography transforms. The pre-distorted images are displayed by the pico projector whose geometrical distortion undoes the pre-distortion such that distortion-free images can be displayed. The proposed algorithm requires a small amount of information stored in the memory and three bilinear interpolations per pixel, thus providing an efficient way to correct nonlinear distortion without additional optical components, such as aspherical lenses.

The main focus of this reported work is to imitate human hand gestures in the presence of the organ matching problem. Towards this aim, a case where the imitator is a fluidic system is adopted, whose dynamics are totally different from the imitatee; that is, a human performing different hand gestures. Proposed is a fluidics formation control of fluid particles where the formations result from the imitation of observed human hand gestures. Fluid body motion is modelled in a particle-based approach by using smoothed particle hydrodynamics, which is the Lagrangian method for the simulation of fluid flows. A shape-based controller for human hand gestures imitation by using the ellipse fitting technique is introduced. With the help of this technique, a novel mapping method between the fluid and ellipse parameters is found. The fluid formation control layer is responsible for assigning the appropriate fluid parameters to the swarm in order to generate observed human hand gestures. Finally, the fluid particle formations are generated and these formations are demonstrated in a simulation environment.

The fractal dimension (FD) is an important feature used for classification and shape recognition. The best method to obtain the FD is the Bouligand-Minkowski method. However, this method is computationally exhaustive because of the use of the distance transform. Presented is a modification in the method and a proposed architecture, suitable for field programmable gate array implementation that enables calculating the FD in a simple and efficient way.

Inter-prediction in high-efficiency video coding (HEVC) provides higher coding efficiency at the expense of increased computational complexity compared with previous video coding standards. Presented is an efficient inter-prediction algorithm based on prediction unit pruning to reduce the computational load. Experimental results show that the proposed algorithm achieves about 11% additional reduction in the inter-mode decision time when applied to a coding unit pruning algorithm.

A wide-field pseudo-heterodyne interference contrast microscope is described, which employs a complementary metal–oxide semiconductor (CMOS) phase-sensitive camera. The use of multiple wells in the camera enables extremely rapid measurement of a full phase field at high resolution and the modulation frequency is not limited by the camera frame rate. The high data acquisition frequency allows the effects of microphonics to be frozen to mitigate the effect of low-frequency disturbance.

The existing thumbnail extraction method generates a thumbnail by reducing the reconstructed frame after a full decoding process. This method is complex and requires considerable time, particularly for higher resolution video. To alleviate these issues, a fast thumbnail extraction method which utilises an intra-prediction mode for high-efficiency video coding (HEVC) is presented. The proposed method reconstructs only the 4×4 boundary pixels needed for a thumbnail image based on the intra-prediction mode. Experimental results indicate that the proposed method significantly reduces the computational complexity and extraction time for a thumbnail image. The visual quality of thumbnail images obtained with this method does not differ significantly from that of images extracted after a full decoding process.

Visual comfort (VC) assessment in stereoscopic three-dimensional (S3D) images using salient object disparity (SOD) is proposed. Through careful observations in S3D displays, it has been discovered that SOD is a key factor that influences VC in S3D images. This is because the human visual system controls the eye's lens to focus on the salient object and adjusts the salient object to the zero parallax plane. Thus, if the salient object is not adjusted to the zero parallax, the S3D image causes severe visual discomfort in the human brain. Based on human depth perception, the disparity of salient object is combined with the maximum disparity as the feature for VC prediction. Experimental results show that the proposed method substantially improves the VC prediction performance in S3D images compared with conventional prediction methods.

To cluster or partition data/signal, expectation-and-maximisation or variational approximation with a mixture model (MM), which is a parametric probability density function represented as a weighted sum of K̂ densities, is often used. However, model selection to find the underlying K̂ is one of the key concerns in MM clustering, since the desired clusters can be obtained only when K̂ is known. A new model selection algorithm to explore K̂ in a Bayesian framework is proposed. The proposed algorithm builds the density of the model order which information criterion such as AIC and BIC or other heuristic algorithms basically fail to reconstruct. In addition, this algorithm reconstructs the density quickly as compared with the time-consuming Monte Carlo simulation using integrated nested Laplace approximation.

A two-stage information fusion scheme is proposed for particle filter. Its main characteristic is that the multisensor data is fused not only in the likelihood model but also in the stage of importance sampling. By incorporating the multisensor data in sequence, the proposal density can be refined step by step; thus high-quality particles and, consequently, a more accurate estimate can be obtained.

A more relaxed condition means that fewer of measurements are needed to ensure the exact sparse recovery from the theoretical aspect. The sufficient condition for the greedy block coordinate descent (GBCD) algorithm is relaxed using the near-orthogonality property. It is also shown that the GBCD algorithm fails when (1/(√K+1)≤δ _K+1 <1).

The energy efficiency of the low-density parity check (LDPC) coded chase combining hybrid automatic repeat request (HARQ) system over the additive white Gaussian noise channel is analysed. The energy efficiency can be improved by increasing the maximum transmission number (K) and reducing the transmit power. The asymptotic energy efficiency corresponding to K = ∞ equals the reciprocal of the threshold signal-to-noise ratio of the forward error correction code, which does not depend on the falling speed of the word error ratio. When the retransmission number is relatively small, optimising equal transmit power and variable transmit power can improve the energy efficiency by up to 10 and 18% respectively, compared with minimum equal power transmission.

Finding the largest code with a given minimum distance is one of the most basic problems in coding theory. A sharpening to the linear programming bound for linear codes in the Lee metric is introduced, which is based on an invariance-type property of Lee compositions of a linear code. Using this property, additional equality constraints are introduced into the linear programming problem, which give a tighter bound for linear Lee codes.

A novel low-phase noise oscillator employing a substrate integrated waveguide (SIW) dual-mode bandpass filter as a frequency stabilisation element within its feedback loop is proposed. The SIW dual-mode filter is implemented based on the use of the two degenerate modes (TE₁₀₂ and TE₂₀₁) and has a larger group delay peak than the conventional SIW cavity resonator. By taking advantage of the large group delay peak, the phase noise of the oscillator has significant improvement. An X-band oscillator with a GaAs HEMT packaged transistor has been designed and measured. At an oscillation frequency of 9.04 GHz, the measured phase noise is −111.9 dBc/Hz at 100 kHz offset. The output power is 6.1 dBm and the figure of merit is −198 dBc/Hz.

A novel compact filtering power divider with high frequency selectivity and a wide stopband is proposed. The bandpass filtering response is implemented by using embedded dual-mode resonators. Source/load cross-coupling is used to improve the frequency selectivity of the passband. In addition, the even- and odd-mode method can be applied to analyse the proposed power divider. The presented filtering power divider has been designed, fabricated and measured. The measured results show reasonable agreement with the simulated ones.

A narrowband, guided-mode resonance (GMR)-based, bandpass filter is designed employing one of the powerful types of memetic algorithms known as the improved shuffled frog leaping algorithm. A high-quality filter with robust performance with a SOI structure is designed with a transmittance of >99.6% at 10.6 μm wavelength. The proposed optimisation/design method is very powerful and is able to optimise complex structures with desired optical responses.

Efficient methods for transducing electromagnetic energy into heat are explained. These methods make use of planar technologies such as microstrip and stripline, along with special ceramic materials developed by Microbiotech S.L. The electromagnetic fundamentals of the presented methods are stated, and the simulated and measured results are presented for each model. The efficiency of the transducers is analysed and compared with the state-of-the-art, and finally the possibility of an array scheme for uniform heating purposes is pointed out.

A carefully designed parallel-plate waveguide measurement system is constructed to measure the reflection phase of a compact electromagnetic bandgap (EBG) structure. A quasi-plane wave is created in the parallel-plate waveguide chamber, which is the basic condition of phase measurement. Two EBG samples with different parameters are measured in this system. The data of the two samples agree well with those of infinite array EBG mode simulations, which proves that this measurement method has good accuracy and reliability. This system reduces the cost of EBG reflection phase measurement and the samples used in the experiment have a much smaller size.

MOSFET-like CNTFETs suffer from band-to-band tunnelling, which in turn causes ambipolar conduction. A new structure based on the gate dielectric constant engineering is proposed. It is observed that the dielectric constant (k) plays an important role in modifying the energy band diagram. Therefore, by selecting suitable values of k over the source, channel and drain regions the tunnelling path at the source–channel interface could be eliminated. Consequently, the ambipolar conduction is suppressed.

An integrated optical device that is able to multiplex modulated signals to an optical orthogonal frequency division multiplexing (OFDM) signal directly in the optical domain is demonstrated. The device operates based on optical inverse fast Fourier transform (IFFT). It is composed of directional couplers, delay lines and a combiner, which was fabricated with silica waveguide technology. The operation was experimentally confirmed by generating 40 and 80 Gbit/s optical OFDM signals.

An eight-wavelength hybrid Si/AlGaInAs/InP microring laser array coupled with an underlying silicon waveguide has been fabricated based on the bonding technique, with the radii varying from 15 to 15.14 μm at a step of 20 nm and with a microring width of 2 μm. Continuous-wave (CW) electrically injected operation is obtained at 280 K, with threshold currents of about 6 mA, channel spacing of about 2 nm and sidemode-suppression ratios above 20 dB.

A method of wavelength beam combining to enhance the total output brightness of vertical external cavity surface emitting lasers (VECSELs) using multiplexed volume Bragg gratings (MVBGs) is presented. The MVBGs are designed to introduce wavelength-specific feedback for the 1020 nm near-infrared wavelength VECSELs to achieve both spectral stabilisation and beam combining effects. In conjunction with the compound cavity arrangement, a combined output of >6 W is achieved with >98% combining efficiency using two VECSELs. The beam combining scheme produces near-diffraction limited output with enhanced spectral density and wavelength-stabilised characteristics.

A new automatic resonant frequency matching technique based on voltage detection by sampling the output transformer loop of the radio-frequency power amplifier (PA) in ultrasonic medical applications is proposed. The proposed automatic resonant frequency matching circuitry is composed of a single-loop coil added to a PA output transformer, voltage divider, diode detector and micro-controller unit for frequency adjustment. Using the proposed circuitry, the resonant frequency shift due to output load variations can be effectively tracked and compensated to provide optimised power transfer to a nonlinear load. For verification, a prototype ultrasonic medical application using a piezoelectric transducer is implemented and examined with silicone oil to check the resonance.

A novel high-efficiency high step-up pulse-width modulation (PWM) resonant converter is proposed. It achieves high-voltage step-up by connecting the outputs of the boost converter cell and the isolated single-ended PWM resonant converter cell in series, and high efficiency by zero current switching for the diode on the secondary side of the transformer of the single-ended PWM resonant converter cell. The boost converter cell in the proposed converter can not only step-up voltage, but also absorbs the leakage inductor energy during the turn-off interval of the power switch to improve power conversion efficiency. Experimental results of a 100 W 35 V/200 V prototype are presented to verify the analysis results of the proposed converter.

It has been proved that the choice of normalisation gain, which is used to synthesise the predistortion function, has significant influence on the linearity performance of linearised power amplifiers. A one-step method is proposed to extract the optimal normalisation gain. The proposed method is directly derived from physical logics to keep the gain of predistortion function unity. Thus, it has much lower complexity compared to the scanning strategy in the conventional method and has easy implementation. A experimental test shows that the linearity performance of the proposed method is comparable to the optimal normalisation gain and also outperforms other published ones.

A pulse-density-modulated (PDM) flyback-based microinverter with a high-frequency AC (HFAC) link and an active decoupling circuit is proposed. The converter's structure and basic operating principle are analysed. The advantages of utilising an active decoupling circuit and the demand for using pulse-density modulation are explained. The operation of the proposed microinverter has been verified by experiment.

Geometric dilution of precision (GDOP) is a measurement of localisation accuracy in bearing-only passive location. However, the current GDOP expression is available only for two dimensions; while GDOP is widely adopted in spatial application, consequently, the explicit GDOP expression in three-dimensional (3D) space using the Cauchy-Binet formula is derived. The derived GDOP expression is used in two examples, which shows that the localisation accuracy is affected by both the number of sensors and the observation geometry.

A novel method employing two-dimensional (2D) unitary ESPRIT and a spatial smoothing technique is proposed for super-resolution ISAR imaging. The spatial smoothing technique is employed to resolve the contradiction that 2D unitary ESPRIT requests multiple snapshots; however, in radar imaging applications only one snapshot of the radar data are available. Experimental results are presented to verify the high resolution and feasibility of the proposed imaging algorithm in ISAR imaging.

A direct way to measure the electrical value of capacitors embedded in a circuit using a ring-oscillator is presented. A calibration system ensures the robustness of the measurement process against temperature, power supply and process variations. Silicon results show the ability of the system to measure robustly a large range of small capacitors. The system also provides a noise immunity as the system provides a digital signature of the capacitor value.

In p-GaN gate AlGaN/GaN power devices, the p-GaN etching process to define the gate region is critical to device performance. In some cases, the remained p-GaN in the ungated region can exist as a result of under-etching, and can act like a charge reservoir in the p-GaN gate power device. In this reported work, the effect of this remained p-GaN layer is investigated for the first time. The main effect of the thick remained p-GaN on the ungated region is the increments of turn-off energy loss (E _off) in spite of the similar turn-on energy loss (E _on). This is related to the hole trapping in the remained p-GaN region which is observed by the change of turn-off gate voltages.

For echo cancellation, a memory improved proportionate M-estimate affine projection algorithm (MIP-MAPA) is proposed in which the M-estimator-based cost function containing the weighting matrix is minimised as the cost function and an efficient recursive implementation is applied according to the memory of the previous step-size control matrices. A test on echo cancellations under an impulsive noise environment showed that the proposed MIP-MAPA achieves improvement in steady-state misalignment compared with existing algorithms.

The binary mask approach has been researched to suppress noise and improve speech intelligibility in noisy environments. An algorithm that estimates the binary mask for noise-corrupted speech based on the instantaneous signal-to-noise ratio (SNR) estimation is proposed. The instantaneous SNR estimation is performed by minimising the Bayes risk with a weighted cost function. In the experiments, white noise was used for the training of the SNR estimator and the binary mask estimation was performed for babble, factory, speech-shaped noise. The experimental results show that the proposed method yields substantial improvements in terms of classification accuracy for the binary mask estimation.

An acoustic model adaptation algorithm is proposed for reverberant speech recognition. Inspired by the eigenvoice adaptation framework, multiple acoustic models reflecting various reverberant environments are combined for instantaneous adaptation. Using artificially generated reverberant speech, multiple acoustic models are trained according to multiple reverberation times. The mean vectors of the optimal acoustic model are obtained as a weighted sum of those of multiple acoustic models by using a maximum-likelihood criterion. For effective model combination, reverberant speech is preprocessed. Experiments on English continuous speech recognition tasks in a simulated reverberant environment show that the proposed method performs better than the conventional adaptation techniques.

The error performance of two-hop cooperative communication (CC) systems with transmit antenna selection (TAS) at the source and relay over η − μ fading channels is analysed. The probability density function of the received signal-to-noise ratio and the average bit error rate of two-hop CC systems with a source-destination direct link and TAS at both the source and the relay is derived. The exactness of the analytical expressions is validated by the close agreement of the Monte Carlo simulation results and the analytical results.

To achieve high efficiency and robustness during the node contention process in medium access control (MAC) protocols, a k-round contention resolution (k-CR) scheme is devised. Theoretical analysis of its success probability of selecting only one node p _s and throughput is also presented. Numerical results show that the k-CR scheme can achieve higher p _s than the k-round elimination contention (k-EC) scheme, and higher throughput than the k-EC and IEEE 802.11 DCF schemes in a wireless local area network environment.

A novel probabilistic algorithm for low duty-cycle dynamic (mobile sensors) wireless sensor networks (WSNs) is proposed. The rapid development of smart phones and mobile sensors had fostered a demand for opportunistic encounter-based networking to support mobile social networking and proximity-based gaming. However, the performance of such applications was restricted by a lack of specialised neighbour discovery algorithms. For existing algorithms, even though the probabilistic algorithms have good performance in average discovery latency, when working with a low duty-cycle, the performance is not so encouraging. Deterministic algorithms can provide bound time for the worst discovery latency, but they sacrificed the average discovery latency. In response, a novel prime-set-based neighbour discovery algorithm (PSBA) is proposed. The PSBA is a probabilistic algorithm tailored to low duty-cycle dynamic WSNs. Leveraging on the certainty of the deterministic algorithm and the randomness of the probabilistic algorithm, the inherent long tail of the probabilistic algorithm can be improved. Also, the average discovery latency of the PSBA can be drastically reduced compared to existing algorithms, which precisely meets the requirements in low duty-cycle dynamic WSNs.

Simultaneous wireless information and power transfer (SWIPT) in wireless cooperative networks is studied, where multiple source–destination pairs establish a communication link via an energy harvesting (EH) relay. The relay employs a power splitting (PS) scheme to harvest the energy sent from each source, which it uses to forward the transmissions to the destinations. The total transmission rate maximisation problem is investigated subject to transmit power and EH constraints, which can be divided into each source–destination node’s transmission rate based on a non-cooperative transmission strategy. This problem is addressed to obtain the optimal PS ratio and power allocation, and derive a closed-form solution. Numerical results are also provided to validate the performance of this proposed scheme, which outperforms a scheme with a uniform PS ratio.