방향통계학을 이용한 슬관절 회전축 자료의 분석

김진욱(JinUkKim) 한국체육학회 2009 한국체육학회지 Vol.48 No.6

The sample as unit vector in is called directional data. In order to analyze this sort of directional data we need directional statistics. The purpose of this study is to introduce the exploratory data analysis, estimation, hypothetical test of the directional statistics and to analyze the empirical directional data of instantaneous axis of rotation for knee joint during vertical jump motion. From a series of the axes of rotation acquired by 3-D coordinates of cinematography, the principal axis of rotation of knee joint could be estimated. Direction cosine and spherical coordinates of the principal axis are (-.3037, .4095, .8603) and (114.2˚ , 340.6˚)respectively and 95 % of confidence cone of this axis was elliptical. The angle of major and minor semi-axes of the ellipse to the principal axis of are 7.13˚ , 2.40˚ respectively.

자연과학편 : 오일러 파라미터(Euler parameters)를 이용한 관절회전각도 추정의 안정성

김진욱(JinUkKim) 한국체육학회 2006 한국체육학회지 Vol.45 No.5

There are many methods for describing the human joint motion having more than one rotational degree of freedom in biomechanics, which are the formulae for direction cosines, Euler angles, and helical axis. Based on these methods, many numerical algorithms also have been developed in order to acquire more valid joint kinematical descriptions. The purpose of this study is to verify and validate the possibilities of describing human joint and numerical stability by Euler parameters. For this study, three types of given rotational motion, which are one axis dominant motion of the three sequential rotational movements, were used to analyze the Euler parameters and Gauss random noise with standard deviation 0.01 were added to four artificial marker coordinates. After performing three kinds of simulations for 1000 repetition, validity and stability test were examined. For flexion/extension dominant motion, three Euler angles were well estimated, whereas the other two types of motion estimated Euler angles insufficiently. These results seems to highly relate to marker distribution.

자연과학편 : 다중공선성 상태의 주성분회귀와 능형회귀

김진욱(JinUkKim) 한국체육학회 2006 한국체육학회지 Vol.45 No.4

In regression analysis the alternatives of OLS (ordinary least squares) method for estimating regression coefficients should be considered if there is a severe multicollinearity among explanatory variables. It means that the multicollinearity distorts the results of study. Researcher, therefore, usually use prevailing two alternative methods, which are principal components regression (PCR) and ridge regression (RR). The purpose of this study is to propose the two alternative regression methods in a case of the multicollinearity problem and to give some discussions of the methods. For this study, two regression model were considered. One model contains three explanatory variables, two of three variables of highly correlated, and one response variable, by explanatory variables, which were created under the assumptions of regression analysis. The other model was oxygen consumption analysed by SAS OnlineDoc. From these variables, OLS, PCR and RR estimators were estimated and their variations were examined by means of ridge trace. PCR and RR showed results with satisfaction in a case of multicollinearity of a multiple regression.

슬관절의 굴곡/신전 회전축과 좌표계 축의 일치에 관한 통계적 검정

김진욱(JinUkKim) 한국체육학회 2010 한국체육학회지 Vol.49 No.4

The axes of thigh and shank coordinate system have been assumed each of axis for principal rotations on knee joint motion. The purpose of this study is to verify the flexion/extension axes of thigh and shank coordinate system(anatomical rotation axes) as the mean direction of the instantaneous axes of rotation during pure passive and active of knee flexion/extension motion. Total three ISB coordinate systems, one of thigh and two of shank, were tested. Since the axes of rotation are directional data, directional statistics was used. There was some evidence to reject the hypothesis that mean direction of axes of rotation was the same as the coordinate axis. These results, therefore, suggest that the ISB coordinate systems as reference frames for human movement analysis should be modified.

구관절 중심 추정에 대한 적합법과 변환법의 오차분석

김진욱(JinUkKim) 한국체육학회 2013 한국체육학회지 Vol.52 No.4

There have been many studies on the fitting methods in various area whereas little research has been done on the transformation methods. These two methods are used in biomechanics to estimate the center of rotation in a joint. The purpose of this study is to analyze the statistical properties of transformation methods estimator and compare them to those of fitting methods estimator in a theoretical and an empirical manner. A simulation study for the response of two estimators with two factors, noise and range of motion, was performed. The results showed that the estimator of transformation methods was minimum variance unbiased estimator(MVUE) and had smaller variance and bias than the fitting method. The transformation method, therefore, is superior to the fitting method as regards validity and reliability.

자연과학편 : 관절의 회전축과 회전중심 추정방법 비교

김진욱(JinUkKim) 한국체육학회 2009 한국체육학회지 Vol.48 No.1

In mechanical analysis of human movement the procedure for describing joint motion is crucial. Since this information is a prerequisite to kinetic analysis, it should be estimated with high accuracy. The purpose of this study is to locate the joint center, also referred as the axis of rotation(AoR) and the center of rotation(CoR), during three-dimensional rotational motion. For this purpose two prevailing function methods, Halvorsen, Lesser, & Lundberg(1999) and Gamage & Lasenby(2002), were selected. In this study, extension of their works, three degrees of freedom of simulation were considered to estimate the location vector for joint center. All of the solutions of linear systems of this were estimated by singular value decomposition(SVD). In comparison between two methods for estimating joint center, AoR and CoR, following conclusion could be obtained. For estimating the instantaneous AoR of the spherical joint, the method A was appropriate while the method B was appropriate for CoR of spherical joint fixed with respect to local coordinate system.

자연과학편 : 최소제곱법을 이용한 관절의 각속도 추정

김진욱(JinUkKim) 한국체육학회 2007 한국체육학회지 Vol.46 No.5

A conventional method for the calculation of angular velocity uses Eulerian angles and their angular rates. These rates then are transformed into global or local coordinate system since they are not orthogonal each other. In order to handle the method one must estimate transformation matrix and transform angular rates into orthogonal coordinate system continuously. This might take much computational time and be a complex process. This study, therefore, propose a new method of least squares approach by kinematical equation of relative velocity without estimating transformation matrix. For the purpose, three-dimensional coordinates of artificial markers of two adjacent segments are generated with some noise. The results show that the new method is superior to conventional method under all range of rotational motion and singular situation. This algorithm is stable and has fast calculation time numerically. Furthermore in biomechanical analysis, since almost all three-dimensional coordinates are acquired with respect to global coordinate system, the method proposed by this study is more simple in such a case.

운동역학 : 몬테카를로(Monte Carlo) 방법을 이용한 DLT 계수의 오차분석

김진욱(JinUkKim) 한국체육학회 2005 한국체육학회지 Vol.44 No.5

Usually, the independent variable containing uncertainty has an influence on the dependent variable. Unless the uncertainty is considered, the solution of interested system leads to inaccurate result. The purpose of this study is to explore the error of estimated DLT(direct linear transformation) parameters which are calculated from DLT equations by using the law of propagation of error and Monte Carlo simulation. For this purpose, means and standard deviations on the screen coordinates of 12 samples for 12 control points were measured and used above two methods. From the propagation of error, 11 DLT parameters’ standard deviation were calculated. Then, based on this result, Monte Carlo simulation was performed. The number of 1500 computer generated random number from was used as the independent variables. As entering the random number to DLT equation as the input variables, 1500 times of DLT routine was looped and 1500 samples of each DLT parameter were obtained. Even though the results of this study showed no particular difference between two methods, Monte Carlo simulation method made performing predictable error analysis possible by offering more information of considered system. In the DLT which parameters can be obtained through measurement, it was more understandable for the system to use Monte Carlo simulation.

자연과학편 : 개별적 마커모형에 의한 표본이 회전중심 추정에 미치는 영향

김진욱(JinUkKim) 한국체육학회 2015 한국체육학회지 Vol.54 No.2

STA, soft tissue artefact, takes a huge proportion of errors in photogrammetry. In order to estimate the center of rotation of a human joint with markers having STA, one should use the optimal estimation method in various situations. The purpose of this study is to compare two method of estimating the center of rotation, fitting and transformation methods, given factors of STA magnitude and of functional movements. For combinations of three levels of STA magnitude and three levels of functional movement, each of independent 1000 simulation experiments was performed. The results showed that the bias and the variance of estimator of transformation method had smaller amount than that of fitting method. Therefor, it can be concluded that the transformation method was superior to the fitting method.

운동역학 : 스포츠기술 교정을 위한 운동역학적 분석시스템 개발(자전거 페달밟는 동작을 대상으로)

김기형(KiHyoungKim),유경석(KyoungSugYoo),김형수(HyungSooKim),이현섭(HyunSeobLee),김진욱(JinUkKim),김성일(SungIlKim),김민기(MinGiKim),김성호(SungHoKiim) 한국체육학회 1999 한국체육학회지 Vol.38 No.3

The purpose of this study was to develop a motion analysis system for correcting sports technique and the system can perform an image data acquisition and marker tracking automatically. The hardware of the system was composed of CCD camera of SONY and the image data acquisition board of SPECOM. The software was developed to perform image data acquisition, marker detection, calculation of marker centroid and marker tracking automatically. In this study, we analyzed cycle pedaling technique using this system. We compared the motions of the elite cycle athletics with the of nine university students, and the analyzed parameters for the technique were summarized, next. The result were as follows. 1. The orbit of a low limb joint and a pedal crank. The orbits of an ankle joint to the crank axis were different between the two groups ; In elite group, the orbit was located outside of the crank axis, whereas in ordinary group, it was located inside. 2. Ankle variation of a low limb joint for each pedaling phase. Movement angle of ankle was greater in elite group. 3. Variation of angular velocity of crank. At starting phase, there was no difference between the two groups. But during the loading phase, the two groups showed about 0.03 ㎭/sec² difference.