• Journal of Institute of Control, Robotics and Systems
• /
• v.13 no.2
• /
• pp.86-92
• /
• 2007

This paper presents the development of fire-fighting and rescue robot for Outdoor Environment. In the procedure of this development, we follow Target Oriented Design (TOD) which is recognized as the systematic methodology to design a system by specifying the target clearly. For some real fire fighting tasks (e.g. tasks in shopping street and a market), narrow road make it difficult for existing fire engine to access the firing place. On the other hand, for dangerous tasks (e.g. gasoline station and a storehouse) the explosive materials make it impossible for fire-fighters to access the firing place. Moreover, the smoke and the high-temperature caused by fire make fire fighting difficult. In this situation, the solution is to develop the fire-fighting and rescue robot. TOD is performed firstly by analyzing the environment properties of fro place and the demanded tasks and the fire-fighting and rescue robot is manufactured. For safety, the fire fighting robot should be controlled by remote operation to keep the operator away from the fire, and the control system is divided into three parts: the robot controllers, controller for remote operating device and wireless communication system. We have selected and developed appropriate hardware and software for each part of control system with considering TOD. As a result, the fire-fighting robot functions correctly and the performance and usefulness of our control architecture is validated by successfully performing some fire-fighting tasks.

• Journal of the Institute of Electronics Engineers of Korea SC
• /
• v.45 no.3
• /
• pp.42-50
• /
• 2008

We developed a visual inspection system for detecting defective products. Most existing inspection systems are designed to be dedicated to one product, which makes operator spend extra money and time to adopt other products. In this work, we propose a flexible visual inspection system that can inspect various products without any additional major job at a low-cost. The developed system contained image processing algorithm libraries and user-friendly graphic interface for adaptable image-based inspection system. We can find a proper threshold value using the proposed algorithm which uses correlation coefficient between a non-defective product and existing sample images of defective product. And We tested the performance of the proposed algorithm using Otsu's method. The proposed system is applied to a automated inspection line for cellular phone.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.9 no.3
• /
• pp.498-503
• /
• 2005

As important device that decide output load in superannuated thermoelectric power plant which do Turbine Load Set Motor device. This generation of electric power system operated Set Up Range Motor according to Set Up value that operator manufactures by hand circumvolve, and generation of electric power output load derision is consisted by internal action including Motor Therefore, in this research passively output load operated Turbine Motor Drive equipment that can have existing Turbine Load Set Motor Performance developing Digital Drive Unit device design. Also control algorithm implementation and existing Turbine Load Set Motor Drive and connection possibility through designed controlling system to connect basis function that decide development output load with Digital Drive Unit that designed also with existing Motor Drive Unit and can be operated.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.10 no.5
• /
• pp.983-989
• /
• 2006

Current emigration and immigration control inspects passports by the naked eye, registers them by manual input, and compares them with items of database. In this paper, we propose the method to recognize information codes of passports. The proposed passport recognition method extracts character-rows of information codes by applying sobel operator, horizontal smearing, and contour tracking algorithm. The extracted letter-row regions is binarized. After a CDM mask is applied to them in order to recover the individual codes, the individual codes are extracted by applying vertical smearing. The recognizing of individual codes is performed by the RBF network whose hidden layer is applied by ART 2 algorithm and whose learning between the hidden layer and the output layer is applied by a generalized delta learning method. After a photo region is extracted from the reference of the starting point of the extracted character-rows of information codes, that region is verified by the information of luminance, edge, and hue. The verified photo region is certified by the classified features by the ART 2 algorithm. The comparing experiment with real passport images confirmed the good performance of the proposed method.

• Proceedings of the Korean Institute of Electrical and Electronic Material Engineers Conference
• /
• 2001.11b
• /
• pp.176-180
• /
• 2001

There are several factors causing re-work in CMP process such as improper polish time calculation by operator. removal rate decline of the polisher, unstable in-suit pad conditioning, slurry supply module problem and wafer carrier rotation inconsistancy. And conclusively those fundimental reason for the re-work rate increasement is mainly from the cycle time delay between wafer polish and post measurement. Therefore, Wafer thickness measurement in wet condition could be able to remove those improper process conditions which may happen during the process in comparison with the conventional dried wafer measurement system and it can be able to reduce the CMP process cycle time. CMP scrap reduction by overpolish, re-work rate reduction, thickness control efficiency also can be easily achieved. CMP Equipment manufacturer also trying to develop integrated system which has multi-head & platen, cleaner, pre & post thickness measure and even control the polish time from the calculated removal rate of each polishing head by software. CMP re-work problem such as over & under polish by target thickness may result in the cycle time delay. By reducing those inefficient factors during the process and establish of the automatic process control, CLC system need to be adopted to maximize the process performance. Wafer to Wafer Polish Time Feed Back Control by measuring the wafer right after the polish shorten the polish time calculation for the next wafer and it lead to the perfact Post CMP target thickness control capability. By Monitoring all of the processed the wafer, CMP process will also be stabilize itself.

• Journal of KIISE:Databases
• /
• v.32 no.5
• /
• pp.487-497
• /
• 2005

A range mosaic query returns distribution of data within the query region as a pattern of mosaic, whereas a range aggregate query returns a single aggregate value of data within the query region. The range mosaic query divides a query region by a multi-dimensional grid, and calculates aggregate values of grid cells. In this paper, we propose a new type of query, range mosaic query and a new operator, mosaic-by, with which the range mosaic queries can be represented. In addition, we suggest efficient algorithms for processing range mosaic queries using an aggregate R-tree. The algorithm that we present computes aggregate results of every mosaic grid cell by one time traversal of the aggregate R-tree, and efficiently executes the queries with only a small number of node accesses by using the aggregate values of the aggregate R-tree. Our experimental study shows that the range mosaic query algorithm is reliable in terms of performance for several synthetic datasets and a real-world dataset.

• Journal of KIISE:Databases
• /
• v.27 no.1
• /
• pp.113-128
• /
• 2000

The adaptation of spatial view is to update materialized view objects when apatial view is redefined. There are tow kinds of adaptation : incremental updates and recomputation. The incremental update changes related view objects and it is more efficient than the recomputation which evaluates redefined view defining query because spatial view is defined by spatial query including high cost spatial operator. This paper proposes the several incremental update methods according to the types of changing the definition of a spatial view. There are two kinds of incremental view adaptation : the method of using only the existing view objects and the view derivation relationship between view objects and their sources. This incremental update is achieved by updating the current materialized view objects or by inserting new materialized view objects. Spatial view adapter is implemented and tested on top of the object oriented geographic information system. This paper evaluates performance between the recomputation and incremental update method through real data.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.36 no.9C
• /
• pp.546-554
• /
• 2011

Near targets in the infrared (IR) images have the steady feature for inner region and the transient feature for the boundary region. Based on these features, this paper proposes a new method to extract the fine target shape of near targets in the IR images. First, we detect the boundary region of the candidate targets using the local variance weighted information entropy (WIE) of the original images. And then, a coarse target region can be estimated based on the labeling of the boundary region. For the coarse target region, we use the opening filter with an adaptive structure element to extract the fine target shape. The decision of the adaptive structure element size is optimized for the width information of target boundary by calculating the average WIE in the enlarged windows. The experimental results show that a proposed method has better extraction performance than the previous threshold algorithms.

• KIPS Transactions on Software and Data Engineering
• /
• v.6 no.12
• /
• pp.573-580
• /
• 2017

Deep Neural Network (DNN) is a large layered neural network which is consisted of a number of layers of non-linear units. Deep Learning which represented as DNN has been applied very successfully in various applications. However, many issues in DNN have been identified through past researches. Among these issues, generalization is the most well-known problem. A Recent study, Dropout, successfully addressed this problem. Also, Dropout plays a role as noise, and so it helps to learn robust feature during learning in DNN such as Denoising AutoEncoder. However, because of a large computations required in Dropout, training takes a lot of time. Since Dropout keeps changing an inter-layer representation during the training session, the learning rates should be small, which makes training time longer. In this paper, using mutation operation, we reduce computation and improve generalization performance compared with Dropout. Also, we experimented proposed method to compare with Dropout method and showed that our method is superior to the Dropout one.

• Journal of Microbiology and Biotechnology
• /
• v.26 no.12
• /
• pp.2051-2059
• /
• 2016

To date, there has been no employment of a magnetoelastic (ME) biosensor method to detect Salmonella enterica serovar Typhimurium in soil. The ME biosensor method needs to be investigated and modified for its successful performance. The filtration method, cation-exchange resin method, and combinations of both methods were employed for the extraction of S. Typhimurium from soil. The number of S. Typhimurium and the resonant frequency shift of the ME sensor were then compared using a brilliant green sulfa agar plate and an HP 8751A network analyzer. A blocking study was performed using bovine serum albumin (BSA), polyethylene glycol (PEG), and casein powder suspension. Finally, the modified ME biosensor method was performed to detect S. Typhimurium in soil. The number of S. Typhimurium was significantly decreased from 7.10 log CFU/soil to 4.45-4.72 log CFU/soil after introduction of the cation-exchange resin method. The greatest resonant frequency shift of the measurement sensor was found when employing centrifugation and filtration procedures. The resonant frequency shift of the PEG-blocked measurement sensor was $3,219{\pm}755Hz$, which was significantly greater than those of the BSA- and casein-blocked ME sensor. The optimum concentration of PEG was determined to be 1.0 mg/ml after considering the resonant shift and economic issue. Finally, the modified ME biosensor method was able to detect S. Typhimurium in soil in a dose-response manner. Although these modifications of the ME biosensor method sacrificed some advantages, such as cost, time effectiveness, and operator friendliness, this study demonstrated a novel approach of the ME biosensor method to detect S. Typhimurium in soil.