• Journal of Astronomy and Space Sciences
• /
• v.22 no.3
• /
• pp.273-282
• /
• 2005

A star sensor daytime observation model is developed in order to test the performance of the star sensor useful for daylight application. The centroid errors of the star sensor in the day time application are computed by using the model. The standard atmospheric model (LOWTRAN7) is utilized to calculate the physical quantities of the daylight atmospheric environments where the star sensor is immersed. This observation model takes the separation angles between the sun and star, the centroid algorithm and the various system specifications of the star sensor into the account. The developed star sensor model will provide more realistic measurement errors in estimating the performance of the attitude determination from the vector observations.

• Journal of IKEEE
• /
• v.17 no.3
• /
• pp.254-259
• /
• 2013

The choice of labeled data in semi-supervised learning algorithm can result in effects on the performance of the resultant classifier. In order to select labeled data required for the training of a semi-supervised learning algorithm, VCNN(Vector Centroid Neural Network) is proposed in this paper. The proposed selection method of label data is evaluated on UCI dataset and caltech dataset. Experiments and results show that the proposed selection method outperforms conventional methods in terms of classification accuracy and minimum error rate.

• Journal of computational fluids engineering
• /
• v.14 no.1
• /
• pp.18-23
• /
• 2009

A novel adaptive mesh refinement(AMR) strategy based on the Moment-of-Fluid(MOF) method for volume-tracking dynamic interface computation is presented. The Moment-of-Fluid method is a new interface reconstruction and volume advection method using volume fraction as well as material centroid. The adaptive mesh refinement is performed based on the error indicator, the deviation of the actual centroid obtained by interface reconstruction from the reference centroids given by moment advection process. Using the AMR-MOF method, the accuracy of volume-tracking computation with evolving interfaces is improved significantly compared to other published results.

• Proceedings of the Korean Society of Machine Tool Engineers Conference
• /
• 2002.10a
• /
• pp.393-398
• /
• 2002

This study proposes a method to reduce error scanning data of laser scanner. The method co of 3 stages. First, there is an error indu difference of the distance between the prob the object. It is possible to reduce the e planning a scanning strategy: object settin path. Second, the scan data loss of the tooli affects calculating the tooling ball con z-direction compensation is given to calculat accurate registration points. Third, three p used to determine a transformation matrix on frame. As merging, the maximum error usually on the third tooling ball in the conven method, which select a point among three po randomly. We find the centroid of 3 points apply it to determine a new transformation mat

• 제어로봇시스템학회:학술대회논문집
• /
• 1991.10a
• /
• pp.642-645
• /
• 1991

This paper presents the results of research on hardware and software of the landmark tracking system to the positions of moving robot in real time. The landmark tracking system is composed of CCD camera, landmark, strobo system and image processing board. The algorithm calculates the position and direction by using the coordinate transformation fomula after calculating the centroid and rotation angle of landmark at fixed position using the image data. The experiment is performed with landmark tracking system is loaded on xyz-table. XYZ-table is used for identifying the true position in our experiment. The results shows that this system has high performance with maxima error of .+-.1 pixels.

• Journal of the Society of Naval Architects of Korea
• /
• v.54 no.5
• /
• pp.387-392
• /
• 2017

There is a constant demand for hull variation related to ship design. Various input variables are generally given to achieve the objective functions assigned by each variation process. When dealing with geometric shapes accompanied by nonlinear operations during the variation process, vague relationships or uncertainties among input variables are commonly observed. Therefore, it is strongly recommended to identify those uncertainty factors in advance. A method to modify the shape of a closed hull form with a new area and a centroid had been introduced as a new process of hull variation. Since uncertainty between input variables still existed in the method, however, it was not easy for the user to enter the area and the corresponding centroid. To overcome this problem, a method is presented in this paper to provide the feasible region of centroids for a given area. By utilizing the concept and techniques used in the statistics such as the number of samples, probability, margin error, and level of confidence, this method generates the distribution of possible centroids along the regression curve. The result shows that the method helps the user to choose an appropriate input value following his or her design intention.

• Journal of Korean Neurosurgical Society
• /
• v.39 no.4
• /
• pp.256-259
• /
• 2006

Objective : Having a reliable and reproducible measurement technique to measure the sagittal contour in vertebral fractures is paramount to clinical decision making. This study is designed to determine the most reliable measurement technique in osteoporotic vertebral compression fracture. Methods : Fifteen lateral radiographs of thoracic and lumbar fractures were selected and measured on two separate occasions by three spine surgeons using six different measurement techniques [Centroid, Harrison Posterior Tangent Methods and 4 different types of modified Cobb method]. The radiograph quality was assessed and the center beam location was determined. Statistical analysis including ANOVA for repeated measures was carried out using the SAS software [v 8.0]. Results : The inter and intraobserver variance of the Cobb method 4 and Harrison posterior tangent method were significantly lower than the other four methods. The intraobserver correlation coefficients were the most consistent using the Cobb method 4 [0.982]. which was followed by the Harrison posterior tangent [0.953] and Cobb methods 1 [0.874]. The intraobserver agreement [% of repeated measures within 5 degrees of the original measurement] ranged from 42% to 98% for each technique for all three observers, with the Cobb method 4 showing the best agreement [97.8%] followed by the Harrison posterior tangent method [937%]. Conclusion : The Cobb method-4 and Harrison posterior tangent methods, when applied to measuring the kyphosis, are reliable and have a similar small error range. The Cobb method 4 shows the best overall reliability. However, the centroid method and Cobb method using a fractured endplate do not produce an accurate result due to inter and intraobserver differences in determining the baseline.

• Journal of Institute of Control, Robotics and Systems
• /
• v.20 no.7
• /
• pp.774-781
• /
• 2014

This paper presents a fingerprint database construction method for WLAN RSSI (Received Signal Strength Indicator)-based indoor positioning. When RSSI is used for indoor positioning, the fingerprint method can achieve more accurate positioning than trilateration and centroid methods. However, a FD (Fingerprint Database) must be constructed before positioning. This step is a very laborious process. To reduce the drawbacks of the fingerprint method, a radio propagation model-based FD construction method is presented. In this method, an FD can be constructed by a simulator. Experimental results show that the constructed FD-based positioning has a 3.17m (CEP) error. In this paper, a spatial correlation method is presented to estimate the NLOS(Non-Line of Sight) error included in the FD constructed by a simulator. As a result, the NLOS error of the FD is reduced and the performance of the error compensated FD-based positioning is improved. The experimental results show that the enhanced FD-based positioning has a 2.58m (CEP) error that is a reasonable performance for indoor LBS (Location Based Service).

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.45 no.6
• /
• pp.11-17
• /
• 2008

This paper proposes the USN node location recognition method with low power, which is the modified centroid method. The conventional Zigbee node location recognition is used to the three RSSI values from three beacons, respectively. When a person move with node, the moving speed of USN node doesn't fast. Therefore, one among three positions used to the value that it is a previous value. This method doesn't large the error in terms of the exact position. Using hand-off method, we get about 30% power advantage than the conventional centroid method. And Our monitoring system add the function that it is possible to search for the node region by color.

• Publications of The Korean Astronomical Society
• /
• v.9 no.1
• /
• pp.55-68
• /
• 1994

We have reviewed three different techniques to estimate molecular cloud mass, and discussed the uncertainties involved. We found that determination of the most important parameter, the $^{13}CO$ abundance, is not very sensitive to the real LTE conditions, and that any possible error in deriving LTE column density may not introduce an error in the total gas column density, as far as the visual extinction is established for the object cloud. The virial technique always endows the largest mass estimate as there are several uncertainties, even if the cloud is in virial equilibrium. The strong indicator of the cloud perturbation is the centroid velocity dispersion. The mass using CO luminosity is based on the empirical law, but weakly dependent on the virial assumption, thus it still gives a larger mass estimate. The mass discrepancy is likely to be inevitable, and a factor of two or three difference between mass estimates could easily be attributed to the uncertainties mentioned above. The LTE mass estimate may be the most reliable one if we use the relation visual extinction and $^{13}CO$ column density of the object cloud, and the intercept is included.