• Journal of the Korean Institute of Telematics and Electronics
• /
• v.20 no.1
• /
• pp.37-44
• /
• 1983

A class of adaptive control algorithms applied to discrete-time single-input single-output deterministic linear systems is analyzed by using direct model reference adaptive control. Controller parameters are identified with weighted least square Method. And computer simulations reveal that proposed weighted least square method in which the value of depends on the identification error can be used regardless of the sufficient condition of reference input signal.

• The KIPS Transactions:PartA
• /
• v.16A no.2
• /
• pp.61-68
• /
• 2009

We present a fast 3D shape deformation method that achieves smoothly deformed result by approximating a rigid transformation based on moving least squares (MLS). Our modified MLS formulation reduces the computation cost for computing the optimal transformation of each point and still keeps the rigidity of the deformed results. Even complex geometric shapes are easily, intuitively, and interactively deformed by manipulating point and ellipsoidal handles.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2005.05b
• /
• pp.189-193
• /
• 2005

기존의 회귀식을 사용하거나 새로 유도하여 사용하는 경우 모두 일반적으로 회귀분석의 특성을 간과하고 사용하는 경우가 종종 발생한다. 일반적으로 자료들에서 구해진 회귀식은 분명히 독립변수와 종속변수가 구분되어 유도되었음에도 불구하고 이 식을 사용함에 있어서는 간혹 그 구분을 무시하고 역으로 적용하는 경향이 있었다. 그러나, 독립$\cdot$종속변수가 서로 바뀌면, 연직거리의 잔차들로부터 유도되는 기존의 회귀분석에 의하여, 회귀식이 서로 달라지기 때문에 역으로 적용하여서는 안된다. 이를 해결하기 본 연구에서는 상호변수 최소자승 회귀분석법을 제안하였다. 이론적 내용을 검토를 위해 임진강 영평천의 영중수위표 지점의 2001-2003년의 유량측정자료와 수위-유량곡선을 비교 분석하였다. 결론적으로 상호변수 회귀분석을 사용하면, 기존의 잘못 사용해온 관행을 해소할 수 있을 것이다.

• Proceedings of the KIEE Conference
• /
• 2007.07a
• /
• pp.288-289
• /
• 2007

본 논문에서는 통계적 파라미터 불확실성을 포함한 시변 선형 불확정 시스템에 대한 강인 칼만필터링 문제를 고려한다. 최소자승 관점에서 정의된 공칭 칼만필터링 문제의 목적함수를 파라미터 불확실성의 통계적 특성을 이용하여 가용한 측정행렬의 함수로 표현하고, 이로부터 근사화된 선형공간 위로의 사교사영으로 해를 도출할 수 있음을 보인다. 최종적으로 벡터 최소자승 추정기법을 동일하게 적용하여, 순환강인 칼만필터식을 유도하고, 유도된 강인 칼만필터 식이 최근 제안된 강인 최소자승 추정식에 공정잡음 및 측정잡음 분산을 반영한 보완된 형태임을 확인한다.

• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.36 no.5
• /
• pp.438-447
• /
• 2008

This study is focused on reliability based design optimization (RBDO) using moving least squares. A response surface is used to derive a limit-state equation for reliability based design optimization. Response surface method (RSM) with least square method (LSM) or Kriging will be used as a response surface. RSM is fast to make the response surface. On the other hand, RSM has disadvantage to make the response surface of nonlinear equation. Kriging can make the response surface in nonlinear equation precisely but needs considerable amount of computations. The moving least square method (MLSM) is made of both methods (RSM with LSM+Kriging). Numerical results by MLSM are compared with those by LMS in Rosenbrock function and six-hump carmel back function. The RBDO of engine duct of smart UAV is pursued in this paper. It is proved that RBDO is useful tool for aerospace structural optimal design problems.

• Journal of Korean Society for Geospatial Information Science
• /
• v.20 no.3
• /
• pp.49-56
• /
• 2012

Performance analysis on RAIM, which is one of the techniques for monitoring integrity to ensure the reliability of GPS, was conducted in this study. RAIM is such a method which allows its user to monitor integrity in the stand-alone mode. Among the existing RAIM procedures, the representative methods including the RCM (Range Comparison Method), LSRM (Least Square Residual Method), Parity approach and WRAIM (Weighted RAIM) were evaluated, and their performance was analyzed. To validate the performance of the implemented algorithms, fault detection was tried on the clock malfunctioning event of PRN 23 occurred on January 1st, 2004. As a result, it was identified that the LSRM and the WRAIM detected all the faults happened in the event. In the case of RCM, all the states of fault were detected except for the error which occurred as a false alarm at one epoch. Furthermore, simulated biases were added for each satellite to analyze the sensitivity of each algorithm. Consequently, when biases of the 9-13 meters range were simulated for the RCM and LSRM algorithm, all the malfunctions were detected. For the WRAIM method, it could detect range biases greater than 15 meters.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.37 no.4B
• /
• pp.309-316
• /
• 2012

The compensation algorithm for localization using the least-squires method in NLOS(Non Line of Sight) environment is suggested and the performance of the algorithm is analyzed in this paper. In order to improve the localization correction rate of the moving node, 1) the distance value of the moving node that is moving as an constant speed is measured by SDS-TWR(Symmetric Double-Sided Two-Way Ranging); 2) the location of the moving node is measured using the triangulation scheme; 3) the location of the moving node measured in 2) is compensated using the least-squares method. By the experiments in NLOS environment, it is confirmed that the average localization error rates are measured to ${\pm}1m$, ${\pm}0.2m$ and ${\pm}0.1m$ by the triangulation scheme, the Kalman filter and the least-squires method respectively. As a result, we can see that the localization error rate of the suggested algorithm is higher than that of the triangulation as average 86.0% and the Kalman filter as average 16.0% respectively.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.29 no.9C
• /
• pp.1330-1336
• /
• 2004

This paper addresses an adaptive motion vector prediction algorithm to improve the performance of video encoder. The block-based motion vector is characterized by non-stationary local statistics so that the coefficients of LS (Least Squares) based linear motion can be optimized. However, it requires very expensive computational cost. The proposed algorithm using LS approach with spatially varying motion-directed property adaptively controls the coefficients of the motion predictor and reduces the computational cost as well as the motion prediction error. Experimental results show the capability of the proposed algorithm.

• Journal of the Institute of Electronics and Information Engineers
• /
• v.50 no.3
• /
• pp.3-8
• /
• 2013

Conventional MMSE beamforming scheme shows better performance than ZF or MF beamforming, but it does not aim increase of sum rate. This paper proposes sum rate maximizing scheme by representing sum rate as a function of weighted MSE. Proposed scheme increases sum rate iteratively by updating weights and beam whose initial form is conventional MMSE beam. Simulation result shows sum rate of proposed scheme is higher than those of conventional schemes and increases as the number of users increases.

• Proceedings of the Korean Society of Soil and Groundwater Environment Conference
• /
• 2002.09a
• /
• pp.156-160
• /
• 2002

실험적 베리오그램의 모델링에 SA(Simulated Annealing)기법을 이용하였다. 최소 자승법의 해를 구하기 위하여 기존의 상용 프로그램에서 많이 이용되고 있는 반복법에 근거한 방법에 비해서 SA 기법은 초기 가정값에 크게 영향을 받지 않고 일정한 모델 인자의 값을 제시하였다. 임의의 초기 가정값을 입력하여도 충분한 반복 계산을 통하여 목적함수의 값이 광역적 최소값으로 수렴하는 것을 확인할 수 있었다. 베리오그램 모델이 일반적으로 비선형 모델이기 때문에 목적함수의 지역적 최소값으로의 수렴이 문제가 되고 이로 인하여 구해지는 인자의 값이 정확하지 않을 수 있지만 SA 기법을 이용하여 최소 자승법의 해를 구하게 되면 정확한 인자의 값을 구할 수 있음을 확인하였다.