• Journal of Korean Society of Transportation
• /
• v.18 no.6
• /
• pp.89-99
• /
• 2000

To cope with the limits of conventional O-D trip matrix collecting methods, several approaches have been developed. One of them is bilevel Programming method Proposed by Yang(1995), which uses Sensitivity Analysis Based(SAB) algorithm to solve Generalized Least Square(GLS) problem. However, the SAB a1gorithm has revealed two critical short-comings. The first is that when there exists a significant difference between target O-D matrix and true O-D matrix, SAB algorithm may not produce correct solution. This stems from the heavy dependance on the historical O-D information, in special when gravel Patterns are dramatically changed. The second is the assumption of iterative linear approximation to original Problem. Because of the approximation, SAB algorithm has difficulty in converging to Perfect Stackelberg game condition. So as to avoid the Problems. we need a more robust and stable solution method. The main purpose of this Paper is to show the problem of the dependency of Previous models and to Propose an alternative solution method to handle it. The Problem of O-D matrix estimation is intrinsically nonlinear and nonconvex. thus it has multiple solutions. Therefore it is necessary to require a method for searching globa1 solution. In this paper, we develop a solution algorithm combined with genetic algorithm(GA) , which is widely used as probabilistic global searching method To compare the efficiency of the algorithm, SAB algorithm suggested by Yang et al. (1992,1995) is used. From the results of numerical example, the Proposed algorithm is superior to SAB algorithm irrespective of travel patterns.

• Journal of the Korean Institute of Gas
• /
• v.19 no.1
• /
• pp.28-37
• /
• 2015

This study presents the development of a well placement optimization model, combining an artificial neural network, which enables high-speed calculation, with a simulated annealing algorithm. The conventional FDM simulator takes excessive time when used to perform a field scale reservoir simulation. In order to solve this problem, an artificial neural network was applied to the model to allow the simulation to be executed within a short time. Also by using the given result, the optimization method, SA algorithm, was implemented to automatically select the optimal location without taking any subjective experiences into consideration. By comparing the result of the developed model with the eclipse simulator, it was found that the prediction performance of the developed model has become favorable, and the speed of calculation performance has also been improved. Especially, the optimum value was estimated by performing a sensitivity analysis for the cooling rate and the initial temperature, which is the control parameter of SA algorithm. From this result, it was verified that the calculation performance has been improved, as well. Lastly, an optimization for the well placement was performed using the model, and it concluded the optimized place for the well by selecting regions with great productivity.

• Journal of KIISE:Information Networking
• /
• v.31 no.3
• /
• pp.269-279
• /
• 2004

The most important factor influencing the robustness of the Internet Is the end-to-end TCP congestion control. However, the congestion control scheme of TCP Reno, the most popular TCP version on the Internet, employs passive congestion indication. It makes worse the network congestion. Recently, Brakmo and Peterson have proposed a new version of TCP, which is named TCP Vegas, with a fundamentally different congestion control scheme from that of the Reno. Many studies indicate that the Vegas is able to achieve better throughput and higher stability than the Reno. But there are two unfairness problems in Vegas. These problems hinder the spread of the Vegas in current Internet. In this paper, in order to solve these unfairness problems, we propose a new congestion control algorithm called TCP PowerVegas. The existing Vegas depends mainly only on the rtt(round trip time), but the proposed PowerVegas use the new congestion control scheme combined the Information on the rtt with the information on the packet loss. Therefore the PowerVegas performs the congestion control more competitively than the Vegas. Thus, the PowerVegas is able to solve effectively these unfairness problems which the Vegas has experienced. To evaluate the proposed approach, we compare the performance among PowerVegas, Reno and Vegas under same network environment. Using simulation, the PowerVegas is able to achieve better throughput and higher stability than the Reno and is shown to achieve much better fairness than the existing Vegas.

• Journal of Korean Society of Steel Construction
• /
• v.19 no.1
• /
• pp.29-41
• /
• 2007

Structural design, like other complex decision problems, involves many trade-offs among competing criteria. Although mathematical programming models are becoming increasingly realistic, they often have design limitations, that is, there are often relevant issues that cannot be easily captured. From the understanding of these limitations, a decision-support system is developed that can generate some useful alternatives as well as a single optimum value in the optimization of steel frame structures. The alternatives produced using this system are "good" with respect to modeled objectives, and yet are "different," and are often better, with respect to interesting objectives not present in the model. In this study, we created a decision-support system for designing the most cost-effective moment-resisting steel frame structures for resisting lateral loads without compromising overall stability. The proposed approach considers the cost of steel products and the cost of connections within the design process. This system makes use of an optimization formulation, which was modified to generate alternatives of optimum value, which is the result of the trade-off between the number of moment connections and total cost. This trade-off was achieved by reducing the number of moment connections and rearranging them, using the combination of analysis based on the LRFD code and optimization scheme based on genetic algorithms. To evaluate the usefulness of this system, the alternatives were examined with respect to various design aspects.

• Journal of the Institute of Convergence Signal Processing
• /
• v.8 no.4
• /
• pp.237-243
• /
• 2007

In this paper, we propose the Real-time Dead-Pixel Detection and Compensation System for mobile camera and its hardware architecture. The CMOS image sensors as image input devices are becoming popular due to the demand for miniaturized, low-power and cost-effective imaging systems. However a conventional Dead-Pixel Detection Algorithm is disable to detect neighboring dead pixels and it degrades image quality by wrong detection and compensation. To detect dead pixels the proposed system is classifying dead pixels into Hot pixel and Cold pixel. Also, the proposed algorithm is processing line-detector and $5{\times}5$ window-detector consecutively. The line-detector and window-detector can search dead pixels by using one-dimensional(only horizontal) method in low frequency area and two-dimensional(vertical and diagonal) method in high frequency area, respectively. The experimental result shows that it can detect 99% of dead pixels. It was designed in Verilog hardware description language and total gate count is 23K using TSMC 0.25um ASIC library.

• Journal of Korea Society of Industrial Information Systems
• /
• v.12 no.5
• /
• pp.39-46
• /
• 2007

Research on human identification in ubiquitous environment has recently attracted a lot of attention. As one of those research, gait recognition is an efficient method of human identification using physical features of a walking person at a distance. In this paper, we present a human shape extraction and tracking for gait recognition using geodesic active contour models(GACMs) combined with mean shift algorithm The active contour models (ACMs) are very effective to deal with the non-rigid object because of its elastic property. However, they have the limitation that their performance is mainly dependent on the initial curve. To overcome this problem, we combine the mean shift algorithm with the traditional GACMs. The main idea is very simple. Before evolving using level set method, the initial curve in each frame is re-localized near the human region and is resized enough to include the targe region. This mechanism allows for reducing the number of iterations and for handling the large object motion. The proposed system is composed of human region detection and human shape tracking modules. In the human region detection module, the silhouette of a walking person is extracted by background subtraction and morphologic operation. Then human shape are correctly obtained by the GACMs with mean shift algorithm. In experimental results, the proposed method show that it is extracted and tracked efficiently accurate shape for gait recognition.

• Journal of KIISE:Software and Applications
• /
• v.26 no.11
• /
• pp.1314-1323
• /
• 1999

컴퓨터의 사용이 일반화됨에 따라 데이타를 생성하고 수집하는 것이 용이해졌다. 이에 따라 데이타로부터 자동적으로 유용한 지식을 얻는 기술이 필요하게 되었다. 데이타 마이닝에서 얻어진 지식은 정확성과 이해성을 충족해야 한다. 본 논문에서는 데이타 마이닝을 위하여 퍼지 결정트리에 기반한 효율적인 퍼지 규칙을 생성하는 알고리즘을 제안한다. 퍼지 결정트리는 ID3와 C4.5의 이해성과 퍼지이론의 추론과 표현력을 결합한 방법이다. 특히, 퍼지 규칙은 속성 축에 평행하게 판단 경계선을 결정하는 방법으로는 어려운 속성 축에 평행하지 않는 경계선을 갖는 패턴을 효율적으로 분류한다. 제안된 알고리즘은 첫째, 각 속성 데이타의 히스토그램 분석을 통해 적절한 소속함수를 생성한다. 둘째, 주어진 소속함수를 바탕으로 ID3와 C4.5와 유사한 방법으로 퍼지 결정트리를 생성한다. 또한, 유전자 알고리즘을 이용하여 소속함수를 조율한다. IRIS 데이타, Wisconsin breast cancer 데이타, credit screening 데이타 등 벤치마크 데이타들에 대한 실험 결과 제안된 방법이 C4.5 방법을 포함한 다른 방법보다 성능과 규칙의 이해성에서 보다 효율적임을 보인다.Abstract With an extended use of computers, we can easily generate and collect data. There is a need to acquire useful knowledge from data automatically. In data mining the acquired knowledge needs to be both accurate and comprehensible. In this paper, we propose an efficient fuzzy rule generation algorithm based on fuzzy decision tree for data mining. We combine the comprehensibility of rules generated based on decision tree such as ID3 and C4.5 and the expressive power of fuzzy sets. Particularly, fuzzy rules allow us to effectively classify patterns of non-axis-parallel decision boundaries, which are difficult to do using attribute-based classification methods.In our algorithm we first determine an appropriate set of membership functions for each attribute of data using histogram analysis. Given a set of membership functions then we construct a fuzzy decision tree in a similar way to that of ID3 and C4.5. We also apply genetic algorithm to tune the initial set of membership functions. We have experimented our algorithm with several benchmark data sets including the IRIS data, the Wisconsin breast cancer data, and the credit screening data. The experiment results show that our method is more efficient in performance and comprehensibility of rules compared with other methods including C4.5.

• KIPS Transactions on Computer and Communication Systems
• /
• v.3 no.11
• /
• pp.393-402
• /
• 2014

We have experienced an explosive increase of the IoT(Internet of Things) technology based devices including smart phones and the wireless communications. Also the growing power consumption in IEEE 802.11 WLANs(Wireless LANs) driven by these dramatic increases in not only mobile users and but also wireless APs(Access Points). To reduce the power consumption, this paper proposes a wireless AP power saving algorithm, which minimizes the transmission power without decrease the transmission and carrier sense ranges. A wireless AP which is use in our algorithm checks its own original coverage periodically for whether there is a new STA(Station) or not when its transmission power is decreased. Moreover, if there are no signaling message to connect the wireless AP, it changes its operation mode Wake-up to sleep. A Result shows that the proposed AP algorithm can reduce the total power consumption of the wireless AP approximated 18% and 35% compared to the conventional wireless AP with and without the existing power saving algorithm, respectively.

• Journal of Broadcast Engineering
• /
• v.24 no.3
• /
• pp.495-505
• /
• 2019

Typical algorithms of deep learning include CNN(Convolutional Neural Networks), which are mainly used for image recognition, and RNN(Recurrent Neural Networks), which are used mainly for speech recognition and natural language processing. Among them, CNN is able to learn from filters that generate feature maps with algorithms that automatically learn features from data, making it mainstream with excellent performance in image recognition. Since then, various algorithms such as R-CNN and others have appeared in object detection to improve performance of CNN, and algorithms such as YOLO(You Only Look Once) and SSD(Single Shot Multi-box Detector) have been proposed recently. However, since these deep learning-based detection algorithms determine the success of the detection in the still images, stable object tracking and detection in the video requires separate tracking capabilities. Therefore, this paper proposes a method of combining Kalman filters into deep learning-based detection networks for improved object tracking and detection performance in the video. The detection network used YOLO v2, which is capable of real-time processing, and the proposed method resulted in 7.7% IoU performance improvement over the existing YOLO v2 network and 20 fps processing speed in FHD images.

• Korean Journal of Remote Sensing
• /
• v.37 no.3
• /
• pp.475-491
• /
• 2021

With the recent increase in available high-resolution (< ~1 m) satellite SAR images, the demand for precise registration of SAR images is increasing in various fields including change detection. The registration between high-resolution SAR images acquired in different look angle is difficult due to speckle noise and geometric distortion caused by the characteristics of SAR images. In this study, registration is performed in two stages, coarse and fine, using the x-band SAR data imaged at staring spotlight mode of TerraSAR-X. For the coarse registration, a method combining the adaptive sampling method and SAR-SIFT (Scale Invariant Feature Transform) is applied, and three rigid methods (NCC: Normalized Cross Correlation, Phase Congruency-NCC, MI: Mutual Information) and one non-rigid (Gefolki: Geoscience extended Flow Optical Flow Lucas-Kanade Iterative), for the fine registration stage, was performed for performance comparison. The results were compared by using RMSE (Root Mean Square Error) and FSIM (Feature Similarity) index, and all rigid models showed poor results in all image combinations. It is confirmed that the rigid models have a large registration error in the rugged terrain area. As a result of applying the Gefolki algorithm, it was confirmed that the RMSE of Gefolki showed the best result as a 1~3 pixels, and the FSIM index also obtained a higher value than 0.02~0.03 compared to other rigid methods. It was confirmed that the mis-registration due to terrain effect could be sufficiently reduced by the Gefolki algorithm.