• IEIE Transactions on Smart Processing and Computing
• /
• 제6권3호
• /
• pp.210-219
• /
• 2017

This paper proposes a modified min-max algorithm (MMMA) for nonbinary quasi-cyclic low-density parity-check (NB-QC-LDPC) codes and an efficient parallel block-layered decoder architecture corresponding to the algorithm on a graphics processing unit (GPU) platform. The algorithm removes multiplications over the Galois field (GF) in the merger step to reduce decoding latency without any performance loss. The decoding implementation on a GPU for NB-QC-LDPC codes achieves improvements in both flexibility and scalability. To perform the decoding on the GPU, data and memory structures suitable for parallel computing are designed. The implementation results for NB-QC-LDPC codes over GF(32) and GF(64) demonstrate that the parallel block-layered decoding on a GPU accelerates the decoding process to provide a faster decoding runtime, and obtains a higher coding gain under a low $10^{-10}$ bit error rate and low $10^{-7}$ frame error rate, compared to existing methods.

• JSTS:Journal of Semiconductor Technology and Science
• /
• 제17권3호
• /
• pp.473-482
• /
• 2017

Hand gesture recognition is regarded as new Human Computer Interaction (HCI) technologies for the next generation of mobile devices. Previous hand gesture implementation requires a large memory and computation power for hand segmentation, which fails to give real-time interaction with mobile devices to users. Therefore, in this paper, we presents a low latency and memory-efficient hand segmentation architecture for natural hand gesture recognition. To obtain both high memory-efficiency and low latency, we propose a streaming hand contour tracing unit and a fast contour filling unit. As a result, it achieves 7.14 ms latency with only 34.8 KB on-chip memory, which are 1.65 times less latency and 1.68 times less on-chip memory, respectively, compare to the best-in-class.

• 대한기계학회논문집B
• /
• 제22권6호
• /
• pp.788-802
• /
• 1998

A numerical method using dual time stepping and preconditioning procedure for efficient computations of unsteady low speed compressible flow problems is developed. The time-derivative preconditioning method which is valid at low speed flow conditions cannot maintain temporal accuracy because of the modification of the time-derivative term in Navier-Stokes equations. The dual time stepping procedure is incorporated to enable the time accurate computations and this procedure introduces a pseudo-time derivative in addition to the physical time derivative. At a given physical time, an inner iteration can be carried out until a steady state in pseudo-time is achieved. This will effectively yield a time accurate solution. Computational capabilities of the above algorithm are demonstrated through computation of a variety of practical fluid flows and it is shown that the algorithms is efficient in the essentially incompressible flows and low Mach number compressible flows with heat source.

• 한국해양공학회지
• /
• 제16권6호
• /
• pp.55-59
• /
• 2002

Delamination means that cracking occurs on the interface layer between composite laminates. In this paper, to predict the delamination growth in composite laminates subjected to low-velocity impact, the unit load method was introduced, and an eighteen-node 3-D finite element analysis, based on assumed strain mixed formulation, was conducted. Strain energy release rate, necessary to determine the delamination growth, was calculated by using the virtual crack closure technique. The unit load method saves the computation time more than the re-meshing method. The virtual crack closure technique enables the strain energy release rate to be easily calculated, because information of the singular stress field near the crack tip is not required. Hence, the delamination growth in composite laminates that are subjected to low-velocity impact can be efficiently predicted using the above-mentioned methods.

• 대한기계학회논문집
• /
• 제18권7호
• /
• pp.1840-1850
• /
• 1994

Enhancement of numerical algorithms for low speed compressible flow will be considered. Contemporary time-marching algorithm has been widely accepted and applied as the method of choice for transonic, supersonic and hypersonic flows. In the low Mach number regime, time-marching algorithms do not fare as well. When the velocity is small, eigenvalues of the system of compressible equations differ widely so that the system becomes very stiff and the convergence becomes very slow. This characteristic can lead to difficulties in computations of many practical engineering problems. In the present approach, the time-derivative preconditioning method will be used to control the eigenvalue stiffness and to extend computational capabilities over a wide range of flow conditions (from very low Mach number to supersonic flow). Computational capabilities of the above algorithm will be demonstrated through computation of a variety of practical engineering problems.

• 전자공학회논문지B
• /
• 제31B권12호
• /
• pp.51-58
• /
• 1994

Mathematical morphology based on set theory is easy to be implemented in parallel and can be applied to various fields in image analysis. Particularly mophological pattern spectrum can detect critical scales in an image object and quantify various aspects of the shape-size content. In this paper, texture classification using pattern spectrum based on morphological subband decomposition is porposed. The low-low band extracts pattern spectrum features, and the high-low, low-high, and high-high bands extrack the structural information. This approach has the advantages of efficient information extraction, less time-consuming, high accuacy, less computation, and parallel implementation.

• Journal of Electrical Engineering and Technology
• /
• 제7권2호
• /
• pp.163-170
• /
• 2012

This paper presents a DFT (Discrete Fourier Transform) based estimation algorithm for the parameters of a low-frequency oscillating signal. The proposed method estimates the parameters, i.e., the frequency, the damping factor, the mode amplitude, and the phase, by fitting a discrete Fourier spectrum with an exponentially damped cosine function. Parameter estimation algorithms that consider the spectrum leakage of the discrete Fourier spectrum are introduced. The multi-domain mode test functions are tested in order to verify the accuracy and efficiency of the proposed method. The results show that the proposed algorithms are highly applicable to the practical computation of low-frequency parameter estimations based on DFTs.

• 반도체디스플레이기술학회지
• /
• 제19권4호
• /
• pp.88-93
• /
• 2020

Based on the capture-emission energy (CEE) maps of CMOS devices, a physics-informed machine learning model for the bias temperature instability (BTI)-induced threshold voltage shifts and low frequency noise is presented. In order to incorporate physics theories into the machine learning model, the integration of artificial neural network (IANN) is employed for the computation of the threshold voltage shifts and low frequency noise. The model combines the computational efficiency of IANN with the optimal estimation of Gaussian mixture model (GMM) with soft clustering. It enables full lifetime prediction of BTI under various stress and recovery conditions and provides accurate prediction of the dynamic behavior of the original measured data.

• 대한기계학회논문집B
• /
• 제32권2호
• /
• pp.83-91
• /
• 2008

A three-dimensional computation was conducted to understand effects of the low Reynolds number on the performance in a low-speed axial compressor at the design condition. The low Reynolds number can originates from the change of the air density because it decreases along the altitude in the troposphere. The performance of the axial compressor such as the static pressure rise was diminished by the separation on the suction surface with full span and the boundary layer on the hub, which were caused by the low Reynolds number. The total pressure loss at the low Reynolds number was found to be greater than that at the reference Reynolds number at the region from the hub to 85% span. Total pressure loss was scrutinized through three major loss categories in a subsonic axial compressor such as the profile loss, the tip leakage loss and the endwall loss using Denton#s loss model, and the effects of the low Reynolds number on the performance were analyzed in detail.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 2005년도 ICCAS
• /
• pp.1826-1829
• /
• 2005

Recent advancement in wireless communications and electronics has enabled the development of low power sensor network. Wireless sensor network are often used in remote monitoring control applications, health care, security and environmental monitoring. Wireless sensor networks are an emerging technology consisting of small, low-power, and low-cost devices that integrate limited computation, sensing, and radio communication capabilities. Sensor network platform for health care has been designed, fabricated and tested. This system consists of an embedded micro-controller, Radio Frequency (RF) transceiver, power management, I/O expansion, and serial communication (RS-232). The hardware platform uses Atmel ATmega128L 8-bit ultra low power RISC processor with 128KB flash memory as the program memory and 4KB SRAM as the data memory. The radio transceiver (Chipcon CC1000) operates in the ISM band at 433MHz or 916MHz with a maximum data rate of 76.8kbps. Also, the indoor radio range is approximately 20-30m. When many sensors have to communicate with the controller, standard communication interfaces such as Serial Peripheral Interface (SPI) or Integrated Circuit ($I^{2}C$) allow sharing a single communication bus. With its low power, the smallest and low cost design, the wireless sensor network system and wireless sensing electronics to collect health-related information of human vitality and main physiological parameters (ECG, Temperature, Perspiration, Blood Pressure and some more vitality parameters, etc.)