• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.33 no.9C
• /
• pp.652-660
• /
• 2008

Applications running under embedded systems require various device drivers designed for different types and versions of the OS to manage resources effectively. In this paper, an automated device driver generator system which can generate the device drivers to be used in newer versions the target OS is proposed. In the proposed system, the structures of device drivers of specific OS are designed in the templates and stored in a library, and the target device drivers are generated by adding codes to the stored templates. Once device drivers are generated, they are registered into the kernel. The experimental results show that data transfer time has been slightly increased when compared against manually created drivers for TFT-LCD driver, USB interface keyboard/mouse driver, and AC'97 controller drivers. The code size for the generated drivers after compilation has also been increased slightly when compared with manually designed device drivers.

• The Journal of the Korea Contents Association
• /
• v.7 no.6
• /
• pp.22-30
• /
• 2007

Most real word systems such as world economy, stock market, and medical applications, contain a series of dynamic and complex phenomena. One of common methods to understand these systems is to build a model and analyze the behavior of the system. In this paper, we investigated methods for best clustering over time series data. As a first step for clustering, BIC (Bayesian Information Criterion) approximation is used to determine the number of clusters. A search technique to improve clustering efficiency is also suggested by analyzing the relationship between data size and BIC values. For clustering, two methods, model-based and similarity based methods, are analyzed and compared. A number of experiments have been performed to check its validity using real data(stock price). BIC approximation measure has been confirmed that it suggests best number of clusters through experiments provided that the number of data is relatively large. It is also confirmed that the model-based clustering produces more reliable clustering than similarity based ones.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2011.10a
• /
• pp.858-861
• /
• 2011

In this paper, we present a RFID system by using a wireless sensor network. The proposed system is installed in glove for activity monitoring. The RFID reader, to send data by using sensor network platform and RFID tag are small size, the shape of quadrangle, and operate in the frequency of 13.56 MHz. The sensor node can read RFID tags on the various objects used in daily living such as furniture, medicines, and kitchenwares. The sensor node reads the data of RFID tags, it transmits wireless packets to the sink node. The sink node sends the received packet immediately to a server system. The data from each RFID system is collected into a database, and then the data are processed to visualize the measurement of daily living activities of users. We provide a web-based monitoring system, and can see the number of RFID tag readings per day as bar charts. The result of experiments demonstrates that the way we propose can help to check the situation of life for people who live alone.

• Journal of the Korea Society of Computer and Information
• /
• v.14 no.7
• /
• pp.49-56
• /
• 2009

In this paper, we present a daily activity monitoring system by using a wireless sensor network. The proposed system is installed in glove for activity monitoring. The RFID reader, to send data by using sensor network platform and RFID tag are small size, the shape of quadrangle, and operate in the frequency of 13.56 MHz. The sensor node can read RFID tags on the various objects used in daily living such as furniture, medicines, and kitchenwares. The sensor node reads the data of RFID tags, it transmits wireless packets to the sink node. The sink node sends the received packet immediately to a server system. The data from each RFID system is collected into a database, and then the data are processed to visualize the measurement of daily living activities of users. We provide a web-based monitoring system, and can see the number of RFID tag readings per day as bar charts. The result of experiments demonstrates that the way we propose can help to check the situation of life for people who live alone.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.25 no.5
• /
• pp.515-521
• /
• 2015

The SymMerge algorithm is an efficient merging algorithm for input sequences u and v of sizes $\left|u \right|=m$ and $\left|v \right|=n$, $m{\leq}n$. We consider complexity analysis for SymMerge algorithm regarding to the required number of comparisons. The focus of the previous complexity analysis was on finding the values of upper bounds, i.e. showing the asymptotical optimality. In this paper, in a different way from the previous complexity analysis, we show that the overall required number of comparisons for two representative special cases "symmetric case" and "maximum spanning case" can be calculated exactly i.e. the least upper bounds regarding to the required number of comparisons are calculated. Symmerge requires exactly $m\;log\frac{n}{m}+4m-logm-3$ comparisons for symmetric case of sizes $m=2^k,\;n=2^l,l{\geq}k$ of input sequences and exactly $\frac{1}{2}m^2+(m+1)logm-\frac{3}{2}m+2$ comparisons for maximum spanning case of sizes $m=2^k,n=2^m-m$ of input sequences. Additionally we show that the complexity of the Symmerge algorithm regarding to the overall required number of comparisons for these special cases can be defined by recurrence relations.

• Journal of Intelligence and Information Systems
• /
• v.19 no.2
• /
• pp.55-71
• /
• 2013

The prediction model is the main factor affecting the performance of a knowledge-based system for bankruptcy prediction. Earlier studies on prediction modeling have focused on the building of a single best model using statistical and artificial intelligence techniques. However, since the mid-1980s, integration of multiple techniques (hybrid techniques) and, by extension, combinations of the outputs of several models (ensemble techniques) have, according to the experimental results, generally outperformed individual models. An ensemble is a technique that constructs a set of multiple models, combines their outputs, and produces one final prediction. The way in which the outputs of ensemble members are combined is one of the important issues affecting prediction accuracy. A variety of combination schemes have been proposed in order to improve prediction performance in ensembles. Each combination scheme has advantages and limitations, and can be influenced by domain and circumstance. Accordingly, decisions on the most appropriate combination scheme in a given domain and contingency are very difficult. This paper proposes a confidence-based selection approach as part of an ensemble bankruptcy-prediction scheme that can measure unified confidence, even if ensemble members produce different types of continuous-valued outputs. The present experimental results show that when varying the number of models to combine, according to the creation type of ensemble members, the proposed combination method offers the best performance in the ensemble having the largest number of models, even when compared with the methods most often employed in bankruptcy prediction.

• The Journal of the Korea Contents Association
• /
• v.7 no.8
• /
• pp.21-30
• /
• 2007

Due to the growth of embedded systems and the development of semi-conductor and storage devices, digital convergence devises is ever growing. Digital convergence devices are equipments into which various functions such as communication, playing movies and wave files and electronic dictionarys are integrated. Example are portable multimedia players(PMPs), personal digital assistants(PDAs), and smart phones. Therefore, these devices need an efficient file system which manages and controls various types of files. In designing such file systems, the size constraint for small embedded systems as well as performance and compatibility should be taken into account. In this paper, we suggest the partial buffer cache technique. Contrary to the traditional buffer cache, the partial buffer cache is used for only the FAT meta data and write-only data. Simulation results show that we could enhance the write performance more than 30% when the file size is larger than about 100 KBytes.

• Journal of the Korea Society of Computer and Information
• /
• v.16 no.2
• /
• pp.17-23
• /
• 2011

In this paper, daily life stress monitoring system is proposed. The proposed wireless ECG module, reducing the noise and increasing the size of signal, amplification circuit was designed for. Using HRV(Heart Rate Variability), extracted by measuring R-wave, stress diagnostic algorithms to assess the stress of human emotion were developed. For monitoring the activities, the proposed system is consist of small rectangular size for portable and by simple measurement it is possible to measure at any time. Through experiments, the proposed approach to represent user's stress level can be confirmed. Through that, it can see appropriate structure to obtain R-wave for stress assess as well as high resemblance to the clinical electrocardiogram. In this paper, performed experiments was developed nonrestraint measuring and wearable wireless biometric scanner that is able to monitor the heart's electrical activity of everyday life.Using this, the algorithm system, that is able to assess stress index through time-domain and frequency-domain analysis of the front and the rear of performing stress load protocol, was developed,

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.35 no.5A
• /
• pp.496-503
• /
• 2010

The implementation of a RF transceiver for WRAN(Wireless Regional Area Network) system based on IEEE 802.22 standard using Cognitive Radio technology is presented in this paper. A CMOS RF transceiver IC for WRAN system operates in VHF/UHF(54~862MHz) broadband, and employs dual-path direct-conversion configuration and the in-band harmonic distortions are effectively suppressed by exploiting the dual-path direct conversion architecture. For 64QAM(3/4 coding rate) OFDM signal, an EVM of <-31.4dB(2.7%) has been achieved at 10dBm off-chip PA output power and the total chip area with pads is 12.95 mm2. The experimental results show that the proposed CMOS RF transceiver IC has perfect performance for WRAN system based on TDD(Time Division Duplex) mode.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.11 no.1
• /
• pp.14-19
• /
• 2007

As XML is used by standard format for information expression and information exchange in various field, Many Company began to use SVG by user interface or information expression tool of embedded system such as mobile phone based wireless internet. Because SVG has many additional information to keep structure of SVG document exception real data, there is a problem that transfer time of SVG file is so cost for quantity of transmitted data actually. To solve this problem, many researches using compression conception have been conducted for applying to an embedded system. This paper proposes SMETA that can use existing researches using compression concept at once. SMETA divides SVG file to each part that can allocate meaning, and gives semantic metadata to each part without alteration of SVG structure. SMETA can reduce site of transmitted SVG file actually by transmitting specification part of SVG file that metadata does not agree or has unlisted part in an embedded system between an embedded system and server, before transmit whole SVG file. By size of transmitting SVG file is decreasing, transfer time can be shortened accordingly.