• ETRI Journal
• /
• v.30 no.3
• /
• pp.451-460
• /
• 2008

Recently, the power consumption of integrated circuits has been attracting increasing attention. Many techniques have been studied to improve the power efficiency of digital signal processing units such as fast Fourier transform (FFT) processors, which are popularly employed in both traditional research fields, such as satellite communications, and thriving consumer electronics, such as wireless communications. This paper presents solutions based on parallel architectures for high throughput and power efficient FFT cores. Different combinations of hybrid low-power techniques are exploited to reduce power consumption, such as multiplierless units which replace the complex multipliers in FFTs, low-power commutators based on an advanced interconnection, and parallel-pipelined architectures. A number of FFT cores are implemented and evaluated for their power/area performance. The results show that up to 38% and 55% power savings can be achieved by the proposed pipelined FFTs and parallel-pipelined FFTs respectively, compared to the conventional pipelined FFT processor architectures.

• Journal of Power Electronics
• /
• v.8 no.4
• /
• pp.291-300
• /
• 2008

In this paper power system and drive-train for omni-directional autonomous mobile robots with multiple energy storage units are presented. Because in proposed system, which is implemented in soccer robots, the ability of power flow control from of multiple separated energy storage units and speed control for each motor are combined, these robots can be derived by more than one power source. This capability, allow robot to diversify its energy source by employing hybrid power sources. In this research Lithium ion polymer batteries have been used for main and auxiliary energy storage units because of their high power and energy densities. And to protect them against deep discharge, over current and short circuit, a protection circuit was designed. The other parts of our robot power system are DC-DC converters and kicker circuit. The simulation and experimental results show proposed scheme and extracted equations are valid and energy management and speed control can be achieved properly using this method. The filed experiments show robot mobility functions to perform the requested motion is enough and it has a high maneuverability in the field.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.44 no.2 s.314
• /
• pp.11-18
• /
• 2007

Back propagation algorithm based on a gradient-decent method has been widely used to the training of a neural network. However, this algorithm have some problems such as dropping the minimum value in a local area according to an initial value and setting the number of units in a hidden layer when training the neural network. Accordingly, to solve the above-mentioned problems, this paper proposes a genetic algorithm using the training method of the neural network. Thus, the improved genetic algorithm using a new crossover and mutation method is proposed to discriminate 3 bit parity. Experiments confirm that the proposed system is effective for training speed after demonstrating for generation gap, the number of units in the hidden layer, and the number of individuals.

• Proceedings of the IEEK Conference
• /
• 2002.07c
• /
• pp.1772-1775
• /
• 2002

In this paper, we propose a Hangul character recognition method in which new letter components as recognition units are introduced and the MLP (multilayer perceptrons) neural networks are employed for two-step recognition of Hangul. To recognize Hangul character, we divide it into two or three recognition units and extract the direction angle features of them to be fed to the corresponding neural network recognizers. The recognition results of neural network recognizers are combined by another neural network. The experiments were conducted on the Hangul characters from real letter envelopes which are collected in the mail centers in Korea and the results showed that our method performs better than the conventional one.

• Proceedings of the IEEK Conference
• /
• 2003.07e
• /
• pp.2419-2422
• /
• 2003

To improve performance of sentence speech recognition systems, we need to consider perplexity of language model and the number of words of dictionary for increasing vocabulary size. In this paper, we propose a language model of VCCV units for sentence speech recognition. For this, we choose VCCV units as a processing units of language model and compare it with clauses and morphemes. Clauses and morphemes have many vocabulary and high perplexity. But VCCV units have small lexicon size and limited vocabulary. An advantage of VCCV units is low perplexity. This paper made language model using bigram about given text. We calculated perplexity of each language processing unit. The perplexity of VCCV units is lower than morpheme and clause.

• IEIE Transactions on Smart Processing and Computing
• /
• v.2 no.4
• /
• pp.197-202
• /
• 2013

Large gains in speech intelligibility can be obtained using the SNR-based binary mask approach. This approach retains the time-frequency (T-F) units of the mixture signal, where the target signal is stronger than the interference noise (masker) (e.g., SNR > 0 dB), and removes the T-F units, where the interfering noise is dominant. This paper introduces two alternative binary masks based on the distortion constraints to improve the speech intelligibility. The distortion constraints are induced by a gain function for estimating the short-time spectral amplitude. One binary mask is designed to retain the speech underestimated (T-F) units while removing the speech overestimated (T-F)units. The other binary mask is designed to retain the noise overestimated (T-F) units while removing noise underestimated (T-F) units. Listening tests with oracle binary masks were conducted to assess the potential of the two binary masks in improving the intelligibility. The results suggested that the two binary masks based on distortion constraints can provide large gains in intelligibility when applied to noise-corrupted speech.

• Electronics and Telecommunications Trends
• /
• v.33 no.4
• /
• pp.14-22
• /
• 2018

Deep neural networks (DNNs) are widely used in various domains such as speech and image recognition. DNN software frameworks such as Tensorflow and Caffe contributed to the popularity of DNN because of their easy programming environment. In addition, many companies are developing neuromorphic processing units (NPU) such as Tensor Processing Units (TPUs) and Graphical Processing Units (GPUs) to improve the performance of DNN processing. However, there is a large gap between NPUs and DNN software frameworks due to the lack of framework support for various NPUs. A bridge for the gap is a DNN software platform including DNN optimized compilers and DNN libraries. In this paper, we review the technical trends of DNN software platforms.

• Journal of the Institute of Electronics and Information Engineers
• /
• v.50 no.1
• /
• pp.97-101
• /
• 2013

There is a growing need for optical communication systems that convert large volumes of data to optical signals and that accommodate and transmit the signals across long distances. Digital optical communication consists of a master unit (MU) and a slave unit (SU). The MU transmits data to SU using digital optical signals. However, digital optical units that are commercially available or are under development transmit data using two's complement representation. At low input levels, a large number of SSOs (simultaneous switching outputs) are required because of the high rate of bit switching in two's complement, which thereby increases the power noise. This problem reduces the overall system capability because a DSP (digital signal processor) chip (FPGA, CPLD, etc.) cannot be used efficiently and power noise increases. This paper proposes a change from two's complement to a more efficient method that produces less SSO noise and can be applied to existing digital optical units.

• 한국정보디스플레이학회:학술대회논문집
• /
• 2005.07b
• /
• pp.1419-1420
• /
• 2005

We have synthesized a photo patternable blue lightemitting polyfluorene (PF) derivative containing cross-linkable oxetane units. Poly(9,9-bis-(4-octyloxyphenyl)- fluorene-2,7-diyl-alt-9,9-bis-((3-hexyloxy-3'- ethyl)-oxetane)-fluorene-2,7-diyl) has been synthesized by Suzuki coupling polymerization. The relationship between patterning property and several variables such as the intensity of the exposed UV light, the concentrations of additives, has been studied by using optical microscope UV/visible spectroscopy, photoluminescence and scanning electron microscope (SEM). We obtained fine patterns with 10 mm resolution using the polymer film after exposure and development. This patterning method using conjugated polymers can be applicable to make fine pixels for PLEDs and OFETs.

• ETRI Journal
• /
• v.44 no.5
• /
• pp.849-858
• /
• 2022

Dynamic voltage frequency scaling (DVFS) has been widely adopted for runtime power management of various processing units. In the case of neural processing units (NPUs), power management of neural network applications is required to adjust the frequency and voltage every layer to consider the power behavior and performance of each layer. Unfortunately, DVFS is inappropriate for layer-wise run-time power management of NPUs due to the long latency of voltage scaling compared with each layer execution time. Because the frequency scaling is fast enough to keep up with each layer, we propose a layerwise dynamic frequency scaling (DFS) technique for an NPU. Our proposed DFS exploits the highest frequency under the power limit of an NPU for each layer. To determine the highest allowable frequency, we build a power model to predict the power consumption of an NPU based on a real measurement on the fabricated NPU. Our evaluation results show that our proposed DFS improves frame per second (FPS) by 33% and saves energy by 14% on average, compared with DVFS.