• 대한산업공학회지
• /
• 제36권2호
• /
• pp.138-145
• /
• 2010

In this paper, we consider the information processing system, which organizes the collected data to meaningful information when the number of data collected from multiple sources reaches to a predetermined number, and performs any action by processing the collected data, or transmits to other devices or systems. We derive an analytical model to calculate the time until it takes to process information after starting to collect data. Therefore, in order to complete the processing data within certain time constraints, we develop some design criteria to control various parameters of the information processing system. Also, we analyze the discrete time model for packet switching networks considering data with no particular arrival nor drop pattern. We analyze the relationship between the number of required packets and average information processing time through numerical examples. By this, we show that the proposed model is able to design the system to be suitable for user's requirements being complementary the quality of information and the information processing time in the system with time constraints.

• 대한전기학회논문지:시스템및제어부문D
• /
• 제51권5호
• /
• pp.175-181
• /
• 2002

In this paper, fault tolerant digital governor is designed to realize ceaseless controlling and to improve the reliability of control system. Designed digital governor huts duplex I/O module and triplex CPU module and also 2 out of 3 voting algorithm and self diagnostic ability. The Processor module of the system(SIDG-3000) is developed based on 32 Bit industrial microprocessor, which guaranteed high quality of the module and SRAM for data also SRAM for command are separated. The process module also includes inter process communication function and power back up function (SRAM for back-up). System reliability is estimated by using the model of Markov process. It is shown that the reliability of triplex system in mission time can be dramatically improved compared with a single control system Designed digital governor system is applied after modelling of the steam turbine generator system of Buk-Cheju Thermal Power Plant. Simulation is carried out to prove the effectiveness of the designed digital governor system

• 전자공학회논문지S
• /
• 제35S권2호
• /
• pp.1-9
• /
• 1998

ATM was adopted as the switching and multiplexing technique for BISDN which aims at transmitting traffics with various characteristics in a unified network. To construct these ATM networks, the most important aspect is the design of the switching system with high performance and different service capabilities. In this paepr, we analyze the performance of an input and output queueing switch with preemptive priority which is considered to be most suitable for ATM networks. For the analysis of an input queue, we model each input queue as two separate virtual input queues for each priority class and we approximage them asindependent Geom/Geom/1 queues. And we model a virtual HOL queue which consists of HOL cells of all virtual input queues which have the same output address to obtain the mean service time at each virtual input queue. For the analysis of an output quque, we obtain approximately the arrival process into the output queue from the state of the virtual HOL queue. We use a Markov chain method to analyze these two models and obtain the maximum throughput of the switch and the mean queueing delay of cells. and analysis results are compared with simulation to verify that out model yields accurate results.

• 한국경영과학회지
• /
• 제40권3호
• /
• pp.63-72
• /
• 2015

This paper presents the issue that is predicting the movement of an agent in an enclosed space by using the MDP (Markov Decision Process). Recent researches on the optimal path finding are confined to derive the shortest path with the use of deterministic algorithm such as $A^*$ or Dijkstra. On the other hand, this study focuses in predicting the path that the agent chooses to escape the limited space as time passes, with the stochastic method. The MDP reward structure from GIS (Geographic Information System) data contributed this model to a feasible model. This model has been approved to have the high predictability after applied to the route of previous armed red guerilla.

• 한국통신학회논문지
• /
• 제34권5A호
• /
• pp.355-362
• /
• 2009

In real-time communication services, delay constraints are among the most important QoS (Quality of Service) factors. In particular, it is difficult to guarantee the delay requirement over wireless channels, since they exhibit dynamic time-varying behavior and even severe burst-errors during periods of deep fading. Channel throughput may be increased, but at the cost of the additional delays when ARQ (Automatic Repeat Request) schemes are used. For real-time communication services, it is very essential to predict data deliverability. This paper derives the delay distribution and the successful delivery probability within a given delay budget using a priori channel model and a posteriori information from the perspective of queueing theory. The Gilbert-Elliot burst-noise channel is employed as an a Priori channel model, where a two-state Markov-modulated Bernoulli process $(MMBP_2)$ is used. for a posteriori information, the channel parameters, the queue-length and the initial channel state are assumed to be given. The numerical derivation is verified and analyzed via Monte Carlo simulations. This numerical derivation is then applied to a rate control scheme for real-time video transmission, where an optimal encoding rate is determined based on the future channel capacity and the distortion of the reconstructed pictures.

• ETRI Journal
• /
• 제45권5호
• /
• pp.822-835
• /
• 2023

Recent semantic segmentation frameworks usually combine low-level and high-level context information to achieve improved performance. In addition, postlevel context information is also considered. In this study, we present a Context ReFinement Network (CRFNet) and its training method to improve the semantic predictions of segmentation models of the encoder-decoder structure. Our study is based on postprocessing, which directly considers the relationship between spatially neighboring pixels of a label map, such as Markov and conditional random fields. CRFNet comprises two modules: a refiner and a combiner that, respectively, refine the context information from the output features of the conventional semantic segmentation network model and combine the refined features with the intermediate features from the decoding process of the segmentation model to produce the final output. To train CRFNet to refine the semantic predictions more accurately, we proposed a sequential training scheme. Using various backbone networks (ENet, ERFNet, and HyperSeg), we extensively evaluated our model on three large-scale, real-world datasets to demonstrate the effectiveness of our approach.

• The Journal of the Acoustical Society of Korea
• /
• 제22권1E호
• /
• pp.3-10
• /
• 2003

This paper describes two kinds of neural networks for stereoscopic vision, which have been applied to an identification of human speech. In speech recognition based on the stereoscopic vision neural networks (SVNN), the similarities are first obtained by comparing input vocal signals with standard models. They are then given to a dynamic process in which both competitive and cooperative processes are conducted among neighboring similarities. Through the dynamic processes, only one winner neuron is finally detected. In a comparative study, with, the average phoneme recognition accuracy on the two-layered SVNN was 7.7% higher than the Hidden Markov Model (HMM) recognizer with the structure of a single mixture and three states, and the three-layered was 6.6% higher. Therefore, it was noticed that SVNN outperformed the existing HMM recognizer in phoneme recognition.

• 품질경영학회지
• /
• 제25권3호
• /
• pp.22-38
• /
• 1997

In a traditional sequential acceptance sampling plan, it is assumed that the sampled items are independent each other. In this paper, stochastically dependent sequential acceptance sampling plans are dealt when there exists dependency between sampled items. Monte-Calro algorithm is used to find the acceptance and rejection probabilities of a lot. The number of defectives for the test to be accepted and rejected in probability ratio sequential test can be found by using these probabilities. The formula for measures of performance of these sampling plans is developed. Type I and II error probabilities are estimated by simulation. This research can be a, pp.ied to sequential sampling procedures in place of control charts where there is a recognized and necessary dependency during the production processes. Also, dependent multiple acceptance sampling plans can be derived by extending this sequential sampling procedure. As a numerical example, a Markov dependent process model is given, and the characteristics of the sampling plans are examined according to the change of the dependency factor.

• 대한음성학회지:말소리
• /
• 제68권
• /
• pp.49-63
• /
• 2008

This paper describes the development and assessment of a hidden Markov model (HMM) based Tagalog speech synthesis system, where Tagalog is the most widely spoken indigenous language of the Philippines. Several aspects of the design process are discussed here. In order to build the synthesizer a speech database is recorded and phonetically segmented. The constructed speech corpus contains approximately 89 minutes of Tagalog speech organized in 596 spoken utterances. Furthermore, contextual information is determined. The quality of the synthesized speech is assessed by subjective tests employing 25 native Tagalog speakers as respondents. Experimental results show that the new system is able to obtain a 3.29 MOS which indicates that the developed system is able to produce highly intelligible neutral Tagalog speech with stable quality even when a small amount of speech data is used for HMM training.

• 한국경영과학회지
• /
• 제39권4호
• /
• pp.153-165
• /
• 2014

We present a new framework for rental capacity management in which rental capacity is dynamically managed by means of temporary inventory addition/return. While serving customers with its own (native) capacity, the rental firm rents additional rental capacity from an upper echelon rental company so that it can avoid lost sales which may occur when stock is not sufficient, and returns it when stock becomes sufficiently large enough to cope with demands. Formulating the model as a Markov decision process, we investigate a flexible capacity addition/return policy that maximizes the firm's profit with respect to system costs. Numerical study indicates that rental operation with capacity addition/return can be economically favorable over rental operation without capacity expansion/return and can contribute the reduction in the size of native rental capacity.