A method is reported to suppress the sidelobe radiation of a standard pyramidal horn using a thin single-layer metamaterial lens. The metamaterial lens is composed of two identical metallic dipole arrays printed on both sides of a single dielectric layer. By placing the metamaterial lens inside a standard pyramidal horn, the electromagnetic fields over the horn aperture can be modified, resulting in low-sidelobe radiation in both the E- and H-planes of the horn. The measurement of a prototype lens gave a 10 dB reduction of the first sidelobe level in the E-plane, compared with that of the original air-filled pyramidal horn. Other design possibilities with a compensating lens are suggested.

An enhanced perceptron model, referred to as BicNeuron (BN), which is based on the novel concept of contrastive biclusters is investigated. According to this viewpoint, a coherent bicluster (i.e. a subset of data patterns showing high similarity across a particular subset of features) belonging to one class is required to show high discrimination from the nearest patterns of the other class when projected on the same feature set. On each local subspace associated with a different pair of contrastive biclusters a perceptron is induced and the model with highest area under the receiver operating characteristic curve value is selected as the final classifier. Experiments involving the discrimination of epileptic and non-epileptic electroencephalogram (EEG) signals show that the BN approach can be very useful, prevailing significantly on standard and kernel perceptrons in terms of classification accuracy.

A mobile application named HLTapper was designed for Android mobile phones to collect and assess real-time alternating finger tapping movement data such as mean tapping speeds, temporal variations of tapping intervals, horizontal tapping variances, and vertical tapping variances, from built-in sensors while a subject performs the finger tapping movement task according to directions given through the application interface. The results of a controlled experiment (40 subjects including nine subjects with Parkinson disease, 11 healthy age-matched subjects, and 20 healthy young subjects) and a discriminant analysis revealed sensitivity of 85.71% and specificity of 91.42%, on an average, which suggests that HLTapper would be useful for early diagnosis and personalised treatment plan adjustments of patients with Parkinson disease.

An area and energy-efficient physically unclonable function (PUF) based on k-winners-take-all (k-WTA) is presented. Competitive computing dynamics of k-WTA reduces key-generation cell in WTA-based PUF to a single transistor, necessary for area/energy-constrained applications. Differential key generation in the PUF generates stable keys at varying temperature and supply voltage variations. Array design considerations for WTA-PUF are studied. Performance of k-WTA-based PUF is compared against the static random access memory (SRAM)-based PUF.

Cost-effective two-dimensional (2D) discrete cosine transform (DCT) and inverse DCT architectures capable of supporting multiple standards of MPEG, H.264 and VC-1 are presented. The proposed core utilises a 1D core and a transposed memory to achieve a low cost design. Multi-level factor share is implemented in conjunction with distributed arithmetic in a system to enable the sharing of the coefficient matrix circuit in order to reduce hardware costs. The proposed approach employs a time-distribution scheme to enable the simultaneous processing of the first and second dimensions to enhance throughput. A high efficiency of this approach was verified by fabricating a test chip using the TSMC 0.18 μm CMOS process. The architecture has an operating frequency of 200 MHz, and throughput of 800 M-pixels/s with a gate count of 44.5 K.

Since the polar code was discovered by Arikan, many decoding methods have been proposed, one of which is the belief propagation (BP) decoding. Its inherent high parallel level makes it suitable for high-speed applications. A novel BP decoder based on a modified three-by-three kernel matrix is proposed. The corresponding decoding algorithm and frozen bits selection method are presented. The numerical results show that it has a similar error correcting capability compared with the original polar BP decoder. More importantly, the message memory requirement can be reduced from O(N log₂ N) to O(N log₃ N), and the decoding delay can be reduced accordingly.

Non-binary turbo codes promise excellent performance for a wide range of applications and environments, but research in this area is sparse. The performance of a non-binary turbo code defined in GF(4) on a two-way wireless relay channel (TWRC) employing physical-layer network coding, which is affected by impulsive noise is investigated. Simulation results for binary and non-binary turbo codes are presented and validated with an error floor analysis at the relay of the TWRC and it is observed that non-binary turbo codes significantly outperform binary turbo codes at low signal-to-noise ratios.

The theoretical bounds are often used to measure the performance of frequency-hopping sequences (FHSs) in FH multiple access systems. For the periodic partial Hamming correlation properties of FHS sets, two kinds of theoretical lower bound and their improved version were proposed. Some theoretical derivation are carried out, and given the exact relation between the bounds.

A novel multiclass AdaBoost-based extreme learning machine (ELM) ensemble algorithm is proposed, in which the weighted ELM is selected as the basic weak classifier because of its much faster learning speed and much better generalisation performance than traditional support vector machines. AdaBoost acts as an ensemble learning method of a number of weighted ELMs. Then, an ensemble strong classifier is constructed by the weighted majority vote of all the weighted ELMs. Compared with the existing ELM methods, the proposed algorithm solves the problem of how to train the weighted samples by ELM in multiclass classification directly. Experiments on the German Traffic Sign Recognition Benchmark database demonstrate that the proposed algorithm can achieve a high recognition accuracy of 99.12% with a relatively lower computational complexity than many state-of-the-art algorithms.

Power allocation techniques for heterogeneous orthogonal frequency division multiple access networks have been well studied. However, due to the different structure of single carrier frequency division multiple access (SC-FDMA), these techniques are not applicable to networks employing SC-FDMA. Optimal uplink power allocation in multiuser heterogeneous single-input–multiple-output (SIMO) SC-FDMA cellular networks is considered. Users’ sum-rate maximisation problem with per-user transmit power constraint is formulated which is a non-linear non-convex problem. This problem is converted to convex optimisation problem and optimal power allocation scheme is derived. Simulations results demonstrate the performance gain of the optimal scheme over the equal power allocation scheme.

A compact tri-band bandpass filter with high controllability is proposed using two multipath λ/4 resonator-loaded resonators (MRLRs). Separated electric and magnetic coupling (SEMC) is applied into the filter, which introduces multiple transmission zeros (TZs) and enables each bandwidth to be separately controllable. In addition, both the return loss control and centre frequency (CF) control are implemented by changing the physical dimension of MRLRs. Owing to the usage of SEMC, three TZs can be, respectively, allocated either below or above the associated passbands. Moreover, the third passband can be far away from the first passband, e.g. 1.57/5.8 GHz (f ₁/3.7f ₁), when each passband is independently designed. A filter example operating at 1.57/3.5/5.8 GHz (GPS/WiMAX/WLAN) with a size of 0.08λ _g × 0.14λ _g is designed and fabricated. Good agreement between simulation and measurement is achieved to validate its performance.

A new microstrip unbalanced-to-balanced (balun) diplexer based on stub-loaded dual-mode resonators is presented. The proposed balun diplexer primarily consists of two balun filter channels, and both channels are entirely independently designable, thus bringing great flexibility in the balun diplexer design. Besides, owing to the direct connection of two dual-mode balun filters, no additional matching networks are demanded at the common input port, leading to a significant size reduction. For demonstration, a prototype balun diplexer operating at 2.30 and 2.72 GHz is implemented. Both simulated and experimental results are provided in good agreement.

Compact size and tunable band pass filter based on open-loop resonators is proposed. The proposed structure consists of dual-armed microstrip open-loop resonators. A second armed open-loop resonator, which acts as band rejection resonator, is connected directly to the input and output ports to achieve the desired band pass behaviour. The open-loop microstrip resonators are combined with four lumped capacitors to reduce the overall size of the proposed filter. The capacitors are inserted at places where maximum electric field is occurred in order to increase the effective capacitance and reduce the resonant frequency as well. The resonance frequency is translated from 6.75 to 5.3 GHz for WLAN applications. Consequently, the filter size will be more compact and the filter performance will be improved by enhancing the energy stored in the resonator. The proposed design using lumped capacitors represents a reduction in size of more than 35% when compared with conventional filter without using lumped capacitors. A good consistency is achieved between simulated and measured results.

An experimental investigation reports on the effectiveness of multi-path and metal-stacking structure in regard to quality factor improvement for on-chip 8-shaped inductors on standard CMOS process. It is found that the multi-path structure has little positive impact on the 8-shaped inductors, while the metal-stacking structure improves quality factor significantly. For a 0.5 nH 8-shaped inductor with top two layers metal-stacking, a good differential quality factor of ∼15 at 10 GHz and ∼17 at 13 GHz is obtained, which makes the inductor suitable for the design of low-cost and interference-immune over-10 GHz radio-frequency and high-speed integrated circuits.

An error-free wireless transmission of a 9 Gbit/s on–off keying modulated signal as well as a 4 k video signal is demonstrated using resonant tunnelling diodes as active elements in both the transmitter and the receiver. The employed system and the modulation scheme enabling the high data rate are discussed.

A novel quantum-dot cellular automata (QCA) design for the non-restoring binary array divider is presented, which is constructed using the proposed coplanar QCA Exclusive-OR gate and full adder. The results show that proposed designs have better performances than the best existing structures in terms of common design metrics. The presented 3 × 3 and 4 × 4 dividers are able to achieve 65, 75, 45% and 41, 27, 41% reductions in cell count, latency, and area, respectively.

Continuous wave (CW) operation of high-power blue laser diodes (LDs) with polished facets on semi-polar (202̅1̅) gallium nitride (GaN) substrates is demonstrated. Ridge waveguide LDs were fabricated using indium GaN waveguiding layers and GaN cladding layers. At a lasing wavelength of 452 nm, the peak two-facet CW output power from an LD with uncoated facets was 1.71 W at a current of 3 A, corresponding to an optical power density of 32.04 MW/cm² on each facet. The dependence of output power on current did not change with repeated CW measurements, indicating that the polished facets did not degrade under high-power CW operation. These results show that polished facets are a viable alternative to cleaved or etched facets for high-power CW semi-polar LDs.

The voltage–current characteristics of high-power LED are different from the conventional power diodes. Generally, the linear model of LED is used in the controller design for LED drivers. A nonlinear model of LED based on the V–I characteristics of LB-W5SM from Osram opto semiconductors company is built. The nonlinear model of LED is a second-order model. The dynamic performance of buck driver with nonlinear model of high-power LED is analysed. The dynamic equations of the V–I characteristics are deduced with curve-fitting. The comparison simulations of proportional–integral-controlled buck drivers with linear model and nonlinear model are conducted. The dynamic response of the second-order model driver is faster than the linear model.

The non-minimum phase, non-linear nature of the SEPIC (single-ended primary inductor converter) makes it very difficult to get robust operating conditions under load variations. Sliding-mode control method is utilised to design a robust controller for tracking control of SEPIC converters for various applications. The implementation of the method is illustrated with the use pulse-width modulation comparators in the control loop to preclude the risk of modulator saturation and to ease the operations. The performance of the SEPIC converter is investigated through numerical simulation and analogue electronics implementation to show the validity, robustness and feasibility of the proposed controller.

A unified pulse-width modulation (PWM) algorithm is proposed for four-switch three-phase inverter (FSTPI) to reveal the essential relationship among different PWM algorithms. The FSTPI generates only four active vectors and the zero vector is absent, so the zero vector can be synthesised by two opposite active vectors or all four vectors. Effective time for different phase is derived for the proposed unified PWM. By altering the distribution of effective time in a PWM period, various PWM algorithms with different equivalent zero vectors are implemented. In the proposed algorithm, sector identification and trigonometric function calculation are not required and the voltage fluctuation across the two dc-link capacitors is also taken into account. Thus, the desired output voltage can be synthesised simply and precisely. Experimental results confirm the effectiveness of the proposed unified PWM algorithm.

Driver drowsiness is one of main causes of road accidents worldwide and has drawn attentions of automobile industries, government organisations as well as researcher scientists and engineers. A non-contact Doppler radar technique for monitoring driver drowsiness is presented. On the basis of unobtrusive Doppler measurement of the respiratory and heart signals, the statistical analysis shows that there is a consistent dependency of the respiratory depth on drowsiness level of drivers. It shows that the non-contact Doppler radar technique can provide an effective approach to monitor driver drowsiness.

22.2 system and its simplified systems could only perfectly reproduce 3D sound field at the central listening point, but listeners may not always stay at the central listening point. Existing non-central zone sound field reproduction methods are of theoretical defects. To solve this problem, this paper proposes a universal arbitrary listening zone sound field reproduction method basing on sound physical property, experimental results prove that it performs better than traditional methods.

The existence of the bufferbloat and high-bandwidth non-adaptive flows results in the high queuing delay, which affects the performance of interactive applications, and the inter-flow unfairness, respectively. To solve above-mentioned issues, an efficient active queue management algorithm with Controlling Fairness and Delay (CFD) is proposed. Different from most other schemes, CFD can achieve the low queuing delay and inter-flow fairness simultaneously without the queue isolation. Extensive simulation results also prove the effectiveness and feasibility of CFD.

Wireless body area networks (WBANs) is now a well-accepted technology that provides not only efficient and remote health-care services but also targets the integration of future multimedia applications, in which high data rate is compulsory. UWB communication is one of the key solutions that afford such rates, and its extended version, multiband OFDM (MB-OFDM) has recently shown its efficiency with WBANs. However, increasing the data rates may come at the expense of the data reliability which is a critical issue in biomedical applications. Cooperative diversity has proven its ability in offering such robustness to channel errors while avoiding to use multiple antennas. Indeed, a single antenna terminal is sufficient to achieve diversity benefits by sharing its physical resources with cooperating single antenna relays. Aiming at providing a viable compromise between the data rate and accuracy in low-profile UWB body networks, to incorporate a decode-and-forward-based cooperative architecture in MB-OFDM system by resorting to opportunistic relaying approach is proposed, which retains only the best on-body relay to forward the information to the destination, thereby significantly reducing the complexity.

The 3D MIMO beamforming system needs a weighting method to determine the direction of beam whist reducing the interference for other beam areas operating at the same carrier frequency. The challenge is to determine the weights of the 3D MIMO beams to direct each beam towards its cluster of user terminals while placing its nulls at undesired user directions to minimise undesired interference. Therefore, the signal-to-interference-plus-noise ratio should be increased while the interference from the side lobes of the other beams reduced. A weight determining method is presented that constructs horizontal and vertical array weights, respectively, by minimising the mean-square error between the array pattern vector and the unit vector, where the unit vector expresses the desired direction for the array pattern and zero vector expresses the undesired direction. Since the rectangular planar array can be viewed as M linear arrays of N elements, the weight of the M–Nth element can be obtained based on the horizontal and vertical array weights.

A novel wideband down-conversion mixer is proposed for multi-standard wireless applications between 100 MHz and 2.5 GHz. A transconductance stage with two pairs of compensatory transistors is adopted for high linearity. Differing from the ordinary differential multi-gated transistor technique utilised in Gilbert mixer, the design proposed concerns about both g _m″ and g _m compensation of the transconductance stage in order to improve both input-referred third-order intercept point (IIP₃) and 1 dB gain compression point (CP_{1 dB}). Realising in Taiwan Semiconductor Manufacturing Company (TSMC) 180 nm RF CMOS process, the mixer provides a gain of 8.9 dB, a maximum IIP₃ of 5.8 dBm, a maximum CP_{1 dB} of −4 dBm and a minimum noise figure (NF) of 9.6 dB.