• The Journal of the Acoustical Society of Korea
• /
• v.25 no.6
• /
• pp.277-281
• /
• 2006

This paper presents a method for the automatic recognition of pitch accents over syllables. The method that we propose is based on the time-delay recursive neural network (TDRNN). which is a neural network classifier with two different representation of dynamic context: the delayed input nodes allow the representation of an explicit trajectory F0(t) along time. while the recursive nodes provide long-term context information that reflects the characteristics of pitch accentuation in spoken English. We apply the TDRNN to pitch accent recognition in two forms: in the normal TDRNN. all of the prosodic features (pitch. energy, duration) are used as an entire set in a single TDRNN. while in the distributed TDRNN. the network consists of several TDRNNs each taking a single prosodic feature as the input. The final output of the distributed TDRNN is weighted sum of the output of individual TDRNN. We used the Boston Radio News Corpus (BRNC) for the experiments on the speaker-independent pitch accent recognition. π 1e experimental results show that the distributed TDRNN exhibits an average recognition accuracy of 83.64% over both pitch events and non-events.

• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.37 no.4
• /
• pp.350-358
• /
• 2009

This paper describes a result of the guidance and control for re-entry vehicle during TAEM phase. TAEM phase (Terminal Aerial Energy Management phase) has many conditions, such as density, velocity, and so on. Under these conditions, we have optimized trajectory and other states for guidance in TAEM phase. The optimized states consist of 7 variables, down-range, cross range, altitude, velocity, flight path angle, vehicle's azimuth and flight range. We obtained the optimized reference trajectory by DIDO tool, and used feedback linearization with neural network for control re-entry vehicle. By back propagation algorithm, vehicle dynamics is approximated to real one. New command can be decided using the approximated dynamics, delayed command input and plant output, NARMA-L2. The result by this control law shows a good performance of tracking onto the reference trajectory.

• The Journal of the Acoustical Society of Korea
• /
• v.18 no.7
• /
• pp.45-55
• /
• 1999

In this paper, we used the WLMS(Wavelet domain Least Mean Square) adaptive filter based on the wavelet transform to cancel grain noise. Usually, grain noise occurs in changes of the crystalline structure of metals in high temperature environment. It makes the detection of flaw difficult. The WLMS adaptive filtering algorithm establishes the faster convergence rate by orthogonalizaing the input vector of adaptive filter as compared with that of LMS adaptive filtering algorithm in time domain. We implemented the WLMS adaptive filter by using the delayed version of the primary input vector as the reference input vector and then implemented the CA-CFAR(Cell Averaging- Constant False Alarm Rate) threshold estimator. CA-CFAR threshold estimator enables to detect the flaw and back echo signals automatically. Here, we used the output signals of adaptive filter as its input signal. To Cow the statistical characteristic of ultrasonic signals corrupted by grain noise, we performed run test. The results showed that ultrasonic signals are nonstationary signal, that is, signals whose statistical properties vary with time. The performance of each filter is appreciated by the signal-to-noise ratio. After LMS adaptive filtering in time domain, SNR improves to about 2-3㏈ but after WLMS adaptive filtering in wavelet domain, SNR improves to about 4-6㏈.

• International Journal of Automotive Technology
• /
• v.8 no.4
• /
• pp.421-428
• /
• 2007

This paper investigates the combustion and emission characteristics of a compression ignition engine fueled with neat and blended Shell's gas-to-liquid (GTL) fuel, which was derived from natural gas through the Fischer-Tropsch process. The experiments were conducted in a 6-cylinder DI diesel engine with pump timing settings of $6^{\circ},\;9^{\circ}\;and\;12^{\circ}$crank angle before TDC over ECE R49 and US 13-mode cycles separately and compared to a conventional diesel fuel. The results show that GTL exhibited almost the same power and torque output, improved fuel economy and effective thermal efficiency. It was found that GTL displayed lower peak in-cylinder combustion pressure and maximum heat release rate (HRR), the timings of the peak pressure and the maximum HRR were generally delayed, and the combustion durations were almost equivalent for diesel and GTL under the same speed-load condition. The results also indicate that, compared to diesel fuel, GTL blends showed a trend forward decreasing four regulated emissions simultaneously and a higher GTL fraction in blends contributing to further reductions in the emissions. In particular and on average, neat GTL significantly reduced HC, CO, NOx and PM by 16.4%, 17.8%, 18.3% and 32.4%, respectively, for all cases.

• Proceedings of the Korea Society for Simulation Conference
• /
• 2001.10a
• /
• pp.321-324
• /
• 2001

The robot soccer simulation game is a dynamic multi-agent environment. In this paper we suggest a new reinforcement learning approach to each agent's dynamic positioning in such dynamic environment. Reinforcement Beaming is the machine learning in which an agent learns from indirect, delayed reward an optimal policy to choose sequences of actions that produce the greatest cumulative reward. Therefore the reinforcement loaming is different from supervised teaming in the sense that there is no presentation of input-output pairs as training examples. Furthermore, model-free reinforcement loaming algorithms like Q-learning do not require defining or loaming any models of the surrounding environment. Nevertheless it can learn the optimal policy if the agent can visit every state-action pair infinitely. However, the biggest problem of monolithic reinforcement learning is that its straightforward applications do not successfully scale up to more complex environments due to the intractable large space of states. In order to address this problem, we suggest Adaptive Mediation-based Modular Q-Learning(AMMQL) as an improvement of the existing Modular Q-Learning(MQL). While simple modular Q-learning combines the results from each learning module in a fixed way, AMMQL combines them in a more flexible way by assigning different weight to each module according to its contribution to rewards. Therefore in addition to resolving the problem of large state space effectively, AMMQL can show higher adaptability to environmental changes than pure MQL. This paper introduces the concept of AMMQL and presents details of its application into dynamic positioning of robot soccer agents.

• Archives of Plastic Surgery
• /
• v.40 no.4
• /
• pp.304-311
• /
• 2013

With an increasing emphasis on microsurgery skill acquisition through simulated training, the need has been identified for standardised training programmes in microsurgery. We have reviewed microsurgery training courses available across the six continents of the World. Data was collected of relevant published output from PubMed, MEDLINE (Ovid), and EMBASE (Ovid) searches, and from information available on the Internet of up to six established microsurgery course from each of the six continents of the World. Fellowships and courses that concentrate on flap harvesting rather than microsurgical techniques were excluded. We identified 27 centres offering 39 courses. Total course length ranged from 20 hours to 1,950 hours. Student-to-teacher ratios ranged from 2:1 to 8:1. Only two-thirds of courses offered in-vivo animal models. Instructions in microvascular end-to-end and end-to-side anastomoses were common, but peripheral nerve repair or free groin flap transfer were not consistently offered. Methods of assessment ranged from no formal assessment, where an instructor monitored and gave instant feedback, through immediate assessment of patency and critique on quality of repair, to delayed re-assessment of patency after a 12 to 24 hours period. Globally, training in microsurgery is heterogeneous, with variations primarily due to resource and regulation of animal experimentation. Despite some merit to diversity in curricula, there should be a global minimum standard for microsurgery training.

• Korean Journal of Veterinary Research
• /
• v.50 no.3
• /
• pp.247-251
• /
• 2010

This study compared the instrument performance and tissue healing of a steel scalpel with a $CO_2$ laser in an animal urinary bladder surgery model. Landrace and Yorkshire mixed breed pigs were used. Two symmetrical incisions were made in urinary bladder of each pig. One incision was made on the left side of ventral aspect on urinary bladder using a steel scalpel, while the other incision was performed on the right side using a $CO_2$ laser with an 8W output power. Each instrument was evaluated clinically for speed, ease of incision, and extent of bleeding. At 7 and 21 days after initial wounding, each wound was taken for histological observations. The scalpel was an easier instrument to use in the confines of the urinary bladder tissue, compared with the laser. However, there was no significant difference between the two groups. The amount of bleeding was less in the laser group but the time of the incisions was shorter with the scalpel. Scalpel incisions showed complete restoration of the epithelium and muscularis. On the other hand, the laser incisions showed incomplete restoration of the epithelium and muscularis. However, most of wound healing in the laser incisions was accomplished according to the time lapse. Although the scalpel produced less damage to the urinary bladder tissue and was easier to handle than the $CO_2$ laser, it did not provide hemostasis that was helpful for use on highly vascular tissue. The $CO_2$ laser provided good hemostasis, but delayed wound healing. In conclusion, the $CO_2$ laser provided better hemostasis and better surgical field than the scalpel. The $CO_2$ laser was used effectively in urinary bladder incision.

• Journal of Yeungnam Medical Science
• /
• v.31 no.1
• /
• pp.38-42
• /
• 2014

Cardiovascular beriberi is caused by thiamine deficiency and usually presents as high cardiac output failure associated with predominantly right-sided heart failure and rapid recovery after treatment with thiamine. Because of its rarity in developed countries, the diagnosis can often be delayed and missed. We recently experienced a case of cardiovascular beriberi with pulmonary hypertension which successfully treated with thiamine infusion. A 50-year-old man with chronic heavy alcoholics was refered to our department for dyspnea with mental change. Echocardiography showed marked right ventricular (RV) dilatation and flattening of the interventricular septum with a D-shaped deformation of the left ventricle. Moderate tricuspid valve regurgitation was found and estimated RV systolic pressure was 52 mm Hg. Because of his confused mentality and history of chronic alcohol intake, neurological disorder due to thiamine deficiency was suspected and intravenous thiamine was administered and he continuously received a daily dose of 100 mg of thiamine. Follow up echocardiography showed marked reduction of RV dilatation and improvement of a D-shaped deformation of the left ventricle. He finally diagnosed as cardiovascular beriberi on the basis of dramatic response to intravenous thiamine. Thiamine deficiency can cause reversible pulmonary hypertension, and can still be encountered in the clinical setting. Thus high index of suspicion is critically needed for diagnosis.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2007.10a
• /
• pp.180-184
• /
• 2007

For driving a sinusoidal type permanent magnet synchronous motor with a maximum continuous torque, a $120^0$ delayed three phase sinusoidal current input which matched with the absolute rotor position is needed at a stator coil. So, the detection of absolute rotor position is required inevitably. Thus the right angle relationship between stator magnetic field and rotor magnetic field has to be preserved at a stator by this commutation action. The detection of a absolute position for the commutation can be made generally by the output signal analysis of the encoder attached at a motor shaft. This study purposes to design signal processing logic circuits which can detect the absolute position of motor with a modem encoder system and generate the three reference wave for making sinusoidal current input at a stator coil.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.38A no.3
• /
• pp.217-223
• /
• 2013

A digital retrodirective antenna system is easy to modify and upgrade because it can control the phase information of the output signal toward opposite direction to input signal without a priori knowledge of the arrival direction. Due to this advantage, it is possible to perform fast beam tracking. In this paper, a design digital retrosirective array antenna system according to the number of antenna array by using only one digital PLL which finds angle of delayed phase and we test BER performance of this system. When we transmit data at actual communication system, the data modulated onto carrier frequency in order to shift spectrum from base band to another band. So we simulate system considering carrier frequency according to the number of antenna array. As a result, carrier frequency has no impact on the performance.