Approximate multiplier design is an effective technique to improve hardware performance at the cost of accuracy loss. The current approximate multipliers are mostly ASIC-based and are dedicated for one particular application. In contrast, FPGA has been an attractive choice for many applications because of its high performance, reconfigurability, and fast development round. This paper presents a novel methodology for designing approximate multipliers by employing the FPGA-based fabrics (primarily look-up tables and carry chains). The area and latency are significantly reduced by applying approximation on carry results and cutting the carry propagation path in the multiplier. Moreover, we explore higher-order multipliers on architectural space by using our proposed small-size approximate multipliers as elementary modules. For different accuracy-hardware requirements, eight configurations for approximate 8×8 multiplier are discussed. In terms of mean relative error distance (MRED), the error of the proposed 8×8 multiplier is as low as 1.06%. Compared with the exact multiplier, our proposed design can reduce area by 43.66% and power by 24.24%. The critical path latency reduction is up to 29.50%. The proposed multiplier design has a better accuracy-hardware tradeoff than other designs with comparable accuracy. Moreover, image sharpening processing is used to assess the efficiency of approximate multipliers on application.

Plane wave scattering from a circular conducting cylinder and a circular conducting strip has been formulated by equivalent surface currents which are postulated from the scattering geometrical optics (GO) field. Thus derived radiation far fields are found to be the same as those formulated by a conventional physical optics (PO) approximation for both E and H polarizations.

This paper presents a method for many-to-one voice conversion using phonetic posteriorgrams (PPGs) based on an adversarial training of deep neural networks (DNNs). A conventional method for many-to-one VC can learn a mapping function from input acoustic features to target acoustic features through separately trained DNN-based speech recognition and synthesis models. However, 1) the differences among speakers observed in PPGs and 2) an over-smoothing effect of generated acoustic features degrade the converted speech quality. Our method performs a domain-adversarial training of the recognition model for reducing the PPG differences. In addition, it incorporates a generative adversarial network into the training of the synthesis model for alleviating the over-smoothing effect. Unlike the conventional method, ours jointly trains the recognition and synthesis models so that they are optimized for many-to-one VC. Experimental evaluation demonstrates that the proposed method significantly improves the converted speech quality compared with conventional VC methods.

This paper investigates the performance of interleave division multiple access (IDMA)-based random access with various interference canceller structures in order to support massive machine-type communications (mMTC) in the fifth generation (5G) mobile communication system. To support massive connectivity in the uplink, a grant-free and contention-based multiple access scheme is essential to reduce the control signaling overhead and transmission latency. To suppress the packet loss due to collision and to achieve multi-packet reception, non-orthogonal multiple access (NOMA) with interference cancellation at the base station receiver is essential. We use IDMA and compare various interference canceller structures such as the parallel interference canceller (PIC), successive interference canceller (SIC), and their hybrid from the viewpoints of the error rate and decoding delay time. Based on extensive computer simulations, we show that IDMA-based random access is a promising scheme for supporting mMTC and the PIC-SIC hybrid achieves a good tradeoff between the error rate and decoding delay time.

In this paper, we propose a deep model of visual recognition based on hybrid KPCA Network(H-KPCANet), which is based on the combination of one-stage KPCANet and two-stage KPCANet. The proposed model consists of four types of basic components: the input layer, one-stage KPCANet, two-stage KPCANet and the fusion layer. The role of one-stage KPCANet is to calculate the KPCA filters for convolution layer, and two-stage KPCANet is to learn PCA filters in the first stage and KPCA filters in the second stage. After binary quantization mapping and block-wise histogram, the features from two different types of KPCANets are fused in the fusion layer. The final feature of the input image can be achieved by weighted serial combination of the two types of features. The performance of our proposed algorithm is tested on digit recognition and object classification, and the experimental results on visual recognition benchmarks of MNIST and CIFAR-10 validated the performance of the proposed H-KPCANet.

In this letter, a piecewise linear (PWL) sigmoid function approximation based on the statistical distribution probability of the neurons' values in each layer is proposed to improve the network recognition accuracy with only addition circuit. The sigmoid function is first divided into three fixed regions, and then according to the neurons' values distribution probability, the curve in each region is segmented into sub-regions to reduce the approximation error and improve the recognition accuracy. Experiments performed on Xilinx's FPGA-XC7A200T for MNIST and CIFAR-10 datasets show that the proposed method achieves 97.45% recognition accuracy in DNN, 98.42% in CNN on MNIST and 72.22% on CIFAR-10, up to 0.84%, 0.57% and 2.01% higher than other approximation methods with only addition circuit.

This paper describes designing a new pedestrian navigation system using a comprehensive framework called the pedestrian navigation concept reference model (PNCRM). We implement this system as a publicly-available smartphone application and evaluate its positioning performance near Omiya station's western entrance. We also evaluate users' subjective impressions of the system using a questionnaire. In both cases, promising results are obtained, showing that the PNCRM can be used as a tool for designing pedestrian navigation systems, allowing such systems to be created systematically.

In this paper, the artificial noise (AN)-aided multiple-input single-output (MISO) cognitive radio network with simultaneous wireless information and power transfer (SWIPT) is considered, in which the cognitive user adopts the power-splitting (PS) receiver architecture to simultaneously decode information and harvest energy. To support secure communication and facilitate energy harvesting, AN is transmitted with information signal at cognitive base station (CBS). The secrecy energy efficiency (SEE) maximization problem is formulated with the constraints of secrecy rate and harvested energy requirements as well as primary user's interference requirements. However, this challenging problem is non-convex due to the fractional objective function and the coupling between the optimization variables. For tackling the challenging problem, a double-layer iterative optimization algorithm is developed. Specifically, the outer layer invokes a one-dimension search algorithm for the newly introduced tight relaxation variable, while the inner one leverages the Dinkelbach method to make the fractional optimization problem more tractable. Furthermore, closed-form expressions for the power of information signal and AN are obtained. Numerical simulations are conducted to demonstrate the efficiency of our proposed algorithm and the advantages of AN in enhancing the SEE performance.

Analyzing aging-induced delay degradations of ring oscillators (ROs) is an effective way to detect recycled field-programmable gate arrays (FPGAs). However, it requires a large number of RO measurements for all FPGAs before shipping, which increases the measurement costs. We propose a cost-efficient recycled FPGA detection method using a statistical performance characterization technique called virtual probe (VP) based on compressed sensing. The VP technique enables the accurate prediction of the spatial process variation of RO frequencies on a die by using a very small number of sample RO measurements. Using the predicted frequency variation as a supervisor, the machine-learning model classifies target FPGAs as either recycled or fresh. Through experiments conducted using 50 commercial FPGAs, we demonstrate that the proposed method achieves 90% cost reduction for RO measurements while preserving the detection accuracy. Furthermore, a one-class support vector machine algorithm was used to classify target FPGAs with around 94% detection accuracy.

A method of determining emission limits was studied by using the amplitude probability distribution (APD) for low-probability pulsed electromagnetic disturbances due to discharge. The features of this method are 1) without using the previously reported relationship between APD and bit error rate, the limits are derived using the measured impact of a pulsed disturbance on various wireless communication systems having different bandwidths, and 2) disturbances caused by discharge with poor reproducibility are simulated by regularly repeated pulse-modulated sine waves to enable stable evaluation of the communication quality. APD-based limits are determined from the pulse repetition frequency of the simulated disturbance such that the block error rate (BLER) is less than a certain limit in wireless systems that are most sensitive to the pulsed disturbance. In the international standard CISPR 32 regulating electromagnetic disturbance, radiated disturbance due to discharge is excluded from the application of peak detection limits because of its low occurrence probability. In this paper we quantitatively determine appropriate criteria of the probability for the exclusion. Using the method, we measured the impact of low-probability pulsed interference on major wireless systems and found that GSM and Wi-Fi systems were the most sensitive. New APD-based limits were derived on the basis of these findings. The APD-based limits determined by the proposed method enable a valid evaluation of low-occurrence-probability pulsed disturbances without unconditionally excluding the measurement.

This paper presents a comparative loss analysis performed between an LLC converter and a phase-shift converter under the same size conditions using a power supply manufactured for information communications equipment. It is also shown herein that the LLC converter has a much higher ratio of transformer loss to total loss than the phase-shift converter and that the cause is the difference in the number of transformer turns between the two converters. Further, the ON-resistance of the secondary-side rectifier element and the number of transformer primary turns are shown to determine which of the two converters is superior in terms of low loss.

Security in visible-light communication (VLC) has seen increasing importance in recent years. Asymmetrically clipped optical orthogonal frequency division multiplexing (ACO-OFDM) is recognized as one of the most powerful and efficient OFDM techniques. Therefore, it is well suited for use in both uplink and downlink connections. A security protocol based on this technique can facilitate secure uplink and downlink connections. In the present study, a low-complexity physical-layer key-generation encryption method is designed using the ACO-OFDM technique for indoor VLC networks. The security method is contingent on the generation of secret keys from the cyclic prefix OFDM samples positioned in the low-channel impact area to encrypt all signal frames before transmission, throughout the session. Numerical results indicate that the key-generation mechanism should be implemented during downlink data transmission throughout a session period to provide keys for both downlink and uplink connections. In this setup, the handset of the user employs the secret keys generated during downlink data transmission to encrypt its uplink transmission. This setup conserves the battery life of the handset. Additionally, the results indicate that the proposed security method can achieve a zero key mismatch rate with on-the-fly key creation.

Wireless full duplex (FD) communication can double the point-to-point throughput. To fully realize the benefits of the FD technique in wireless local area networks (WLANs), it is important to design the medium access control (MAC) protocols for FD communications. In FD MAC protocols, when a node wins the channel contention and transmits a primary transmission, its destination node can start a secondary transmission triggered by the primary transmission. Each secondary transmitter transmits a data frame even if its backoff timer is not zero. However, the backoff scheme in the FD MAC protocols follows the conventional scheme based on the distributed coordination function (DCF). Therefore, the nodes with FD MAC initialize the contention window (CW) size to minimum CW (CWmin) after their successful secondary transmissions. Therefore, CW initialization in the FD MAC causes further collisions at stations (STAs), which degrades network throughput. This paper proposes a novel backoff scheme for FD MAC protocols. In the proposed scheme, the CW size and backoff timer are not initialized but kept the current value after secondary transmissions. The proposed scheme can mitigate frame collisions at STAs and increase FD-transmission opportunity in the network, and then enhance the throughput significantly. This paper presents comprehensive performance evaluation in simulations, including non-saturation and saturation conditions, and co-existence conditions with legacy half duplex (HD) STAs. For performance analysis, this paper establishes Markov-chain models for the proposed scheme. The analytical results show theoretically that the operation of the proposed scheme enhances network throughput. The simulation results and analytical results show the effectiveness of the proposed scheme.

Green cognitive radio (CR) plays an important role in offering secondary users (SUs) with more spectrum with smaller energy expenditure. However, the energy efficiency (EE) issues associated with green CR for fading channels have not been fully studied. In this paper, we investigate the average EE maximization problem for spectrum-sharing CR in fading channels. Unlike previous studies that considered either the peak or the average transmission power constraints, herein, we considered both of these constraints. Our aim is to maximize the average EE of SU by optimizing the transmission power under the joint peak and average transmit power constraints, the rate constraint of SU and the quality of service (QoS) constraint of primary user (PU). Specifically, the QoS for PU is guaranteed based on either the average interference power constraint or the PU outage constraint. To address the non-convex optimization problem, an iterative optimal power allocation algorithm that can tackle the problem efficiently is proposed. The optimal transmission powers are identified under both of perfect and imperfect channel side information (CSI). Simulations show that our proposed scheme can achieve higher EE over the existing scheme and the EE achieved under perfect CSI is better than that under imperfect CSI.

Matching circuits using LC elements are widely applied to high-frequency circuits such as power amplifier (PA) and low-noise amplifier (LNA). For determining matching condition of multi-stage matching circuits, this paper shows that any multi-stage LC-Ladder matching circuit with resistive termination can be decomposed to the extended L-type matching circuits with resistive termination containing negative elements where the analytical solution exists. The matching conditions of each extended L-type matching circuit are obtained easily from the termination resistances and the design frequency. By synthesizing these simple analysis solutions, it is possible to systematically determine the solution even in a large number of stages (high order) matching circuits.

In the statistic en-route filtering, each report generation node must collect a certain number of endorsements from its neighboring nodes. However, at some point, a node may fail to collect an insufficient number of endorsements since some of its neighboring nodes may have dead batteries. This letter presents a report generation method that can enhance the generation process of sensing reports under such a situation. Simulation results show the effectiveness of the proposed method.

In this paper, we study a radio frequency (RF)-powered backscatter assisted cognitive radio network (CRN), where an eavesdropper exists. This network includes a primary transmitter, a pair of secondary transmitter and receiver, a friendly jammer and an eavesdropper. We assume that the secondary transmitter works in ambient backscatter (AmBack) mode and the friendly jammer works in harvest-then-transmit (HTT) mode, where the primary transmitter serves as energy source. To enhance the physical layer security of the secondary user, the friendly jammer uses its harvested energy from the primary transmitter to transmit jamming noise to the eavesdropper. Furthermore, for maximizing the secrecy rate of secondary user, the optimal time allocation including the energy harvesting and jamming noise transmission phases is obtained. Simulation results verify the superiority of the proposed scheme.

Fast computation of singular value decomposition (SVD) is of great interest in various machine learning tasks. Recently, SVD methods based on randomized linear algebra have shown significant speedup in this regime. For processing large-scale data, computing systems with accelerators like GPUs have become the mainstream approach. In those systems, access to the input data dominates the overall process time; therefore, it is needed to design an out-of-core algorithm to dispatch the computation into accelerators. This paper proposes an accurate two-pass randomized SVD, named block randomized SVD (BRSVD), designed for matrices with a slow-decay singular spectrum that is often observed in image data. BRSVD fully utilizes the power of modern computing system architectures and efficiently processes large-scale data in a parallel and out-of-core fashion. Our experiments show that BRSVD effectively moves the performance bottleneck from data transfer to computation, so that outperforms existing randomized SVD methods in terms of speed with retaining similar accuracy.

In wireless sensor networks (WSN), communication interference and the energy limitation of sensor nodes seriously hamper the network performance such as throughput and network lifetime. In this paper, we focus on the Successive Interference Cancellation (SIC) and Wireless Energy Transmission (WET) technology aiming to design a heuristic power control algorithm and an efficient cross-layer strategy to realize concurrency communication and improve the network throughput, channel utilization ratio and network lifetime. We realize that the challenge of this problem is that joint consideration of communication interference and energy shortage makes the problem model more complicated. To solve the problem efficiently, we adopt link scheduling strategy, time-slice scheduling scheme and energy consumption optimization protocol to construct a cross-layer optimization problem, then use an approximate linearization method to transform it into a linear problem which yields identical optimal value and solve it to obtain the optimal work strategy of wireless charging equipment (WCE). Simulation results show that adopting SIC and WCE can greatly improve communication capability and channel utilization ratio, and increase throughput by 200% to 500% while prolonging the network lifetime.

Some non-acoustic modalities have the ability to reveal certain speech attributes that can be used for synthesizing speech signals without acoustic signals. This study validated the use of ultrasonic Doppler frequency shifts caused by facial movements to implement a silent speech interface system. A 40kHz ultrasonic beam is incident to a speaker's mouth region. The features derived from the demodulated received signals were used to estimate the speech parameters. A nonlinear regression approach was employed in this estimation where the relationship between ultrasonic features and corresponding speech is represented by deep neural networks (DNN). In this study, we investigated the discrepancies between the ultrasonic signals of audible and silent speech to validate the possibility for totally silent communication. Since reference speech signals are not available in silently mouthed ultrasonic signals, a nearest-neighbor search and alignment method was proposed, wherein alignment was achieved by determining the optimal pair of ultrasonic and audible features in the sense of a minimum mean square error criterion. The experimental results showed that the performance of the ultrasonic Doppler-based method was superior to that of EMG-based speech estimation, and was comparable to an image-based method.