• Journal of the Korean Data and Information Science Society
• /
• 제15권2호
• /
• pp.499-506
• /
• 2004

The Unscented Kalman Filter is known to outperform the Extended Kalman Filter for the nonlinear state estimation with a significance advantage that it does not require the computation of Jacobian but EKF has a competitive advantage to the UKF on the performance time. We compare both algorithms on training the artificial neural network. The validation data set is used to estimate parameters which are supposed to result in better fitting for the test data set. Experimental results are presented which indicate the performance of both algorithms.

• 한국정밀공학회지
• /
• 제15권7호
• /
• pp.175-181
• /
• 1998

This paper describes realtime visual tracking system of moving object for three dimensional target using EKF(Extended Kalman Filter). We present a new realtime visual tracking using EKF algorithm and image prediction algorithm. We demonstrate the performance of these tracking algorithm through real experiment. The experimental results show the effectiveness of the EKF algorithm and image prediction algorithm for realtime tracking and estimated state value of filter, predicting the position of moving object to minimize an image processing area, and by reducing the effect by quantization noise of image.

• 한국항해학회지
• /
• 제6권1호
• /
• pp.61-72
• /
• 1982

As it is well-known, in recent years the Kalman filter has been extensively used in the engineering field. The authors tried to apply the extended Kalman filter for optimal estimation of ship's positiion which is fixed by simultaneous visual (or radio) bearings to two known locations. In practical piloting, bearings are generally measured intermittently, so in this case the original Kalman filter can not be applied because of the long sampling time. In this paper, the extended Kalman filter is revised in order to be used in such an unfavorable case, adn the Digital simulation is conducted by using the revised extended Kalman filter under various noise conditiions. Good results have been obtaiend and effectiveness of the proposed filter has been confirmed.

• Journal of Advanced Marine Engineering and Technology
• /
• 제40권3호
• /
• pp.217-222
• /
• 2016

본 논문은 수정된 Gram-Schmidt와 결합한 Potter 또는 Carlson 알고리즘을 가지는 제곱근 확장 칼만 필터에 의한 영구자석 동기전동기의 센서리스 속도 제어에 관한 연구이다. 일반적으로 반올림 오차에 기인하는 칼만 필터의 민감도는 잘 알려진 문제이다. 제곱근 개념과 칼만 필터의 결합은 수치적 성능을 향상할 수 있고 발산과 같은 불안전한 문제를 풀 수 있다. 본 논문에서는 제곱근 확장 칼만 필터의 구현을 위한 설계와 분석을 수행하였다. 설계된 제곱근 확장 칼만 필터의 추정 성능을 입증하기 위해, 고속, 저속, 역 회전, 파라미터 변동, 부하 변동 실험 등 여러 운전 조건 아래에서 실험 결과들을 분석하였다. 또한, 프로그램 코드 크기 및 연산 시간을 비교하였다. 실험적 결과들은 제곱근 확장 칼만 필터에 의한 영구자석 동기전동기의 센서리스 속도 제어가 양호함을 보인다.

• International Journal of Control, Automation, and Systems
• /
• 제3권2호
• /
• pp.173-182
• /
• 2005

In this paper, a target tracking algorithm for tracking maneuvering vehicles is presented. The overall algorithm belongs to the category of an interacting multiple-model (IMM) algorithm used to detect multiple targets using fused information from multiple sensors. First, two kinematic models are derived: a constant velocity model for linear motions, and a constant-speed turn model for curvilinear motions. Fpr the constant-speed turn model, a nonlinear information filter is used in place of the extended Kalman filter. Being equivalent to the Kalman filter (KF) algebraically, the information filter is extended to N-sensor distributed dynamic systems. The model-matched filter used in multi-sensor environments takes the form of a federated nonlinear information filter. In multi-sensor environments, the information-based filter is easier to decentralize, initialize, and fuse than a KF-based filter. In this paper, the structural features and information sharing principle of the federated information filter are discussed. The performance of the suggested algorithm using a Monte Carlo simulation under the two patterns is evaluated.

• 대한전기학회논문지
• /
• 제38권12호
• /
• pp.971-982
• /
• 1989

An integrating position estimation algorithm has been developed for the navigation system of a free-ranging AGV system. The navigation system focused in this research work consists of redundant wheel encoders for the relative position measurement and a vision sensor for the absolute position measurement. A maximum likelihood method and an extended Kalman filter are implemented for enhancing the performance of the position estimator. The maximum likelihood estimator processes noisy, redundant wheel encoder measurements and yields efficient estimates for the AGV motion between each sampling interval. The extended Kalman filter fuses inharmonious positional data from the deadreckoner and the vision sensor and computes the optimal position estimate. The simulation results show that the proposed position estimator solves a generalized estimation problem for locating the vehicle accurately in space.

• Journal of the Korean Society for Industrial and Applied Mathematics
• /
• 제22권2호
• /
• pp.115-123
• /
• 2018

This paper discusses extended Kalman filter method for solving learning problems of multilayered neural networks. A lot of learning algorithms for deep layered network are sincerely suffered from complex computation and slow convergence because of a very large number of free parameters. We consider an efficient learning algorithm for deep neural network. Extended Kalman filter method is applied to parameter estimation of neural network to improve convergence and computation complexity. We discuss how an efficient algorithm should be developed for neural network learning by using Extended Kalman filter.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 2003년도 ICCAS
• /
• pp.978-983
• /
• 2003

In this paper, we propose a fault detection method for extended Kalman filter in decentralized filter structure. To detect a fault, a consistency between filter output and a monitoring signal is tested. State propagators are used to obtain the monitoring signal. However, the output of state propagator increases in magnitude and finally diverges as time runs. To solve such problem, two-propagator method was proposed for linear system. Two propagators are reset by Kalman filter output, alternatively, to avoid divergence. But a test statistics change abruptly at the reset instant in that method. Hence a N-step propagator method is proposed to fix up the problem. In the N-step propagator, only time propagations are performed from k-N+1 step to k step without measurement updates. A test statistics are defined by errors and its covariance between extended Kalman filter and N-step propagator. These fault detection methods are applied to integrated strapdown inertial navigation system (SDINS). By computer simulation, it is shown that the proposed methods detect a fault effectively.

• 제어로봇시스템학회논문지
• /
• 제21권1호
• /
• pp.34-39
• /
• 2015

In this paper, a finite impulse response (FIR) filter is proposed for discrete-time nonlinear systems. The proposed filter is designed by combining the estimate of the perturbation state and nominal state. The perturbation state is estimated by adapting the optimal time-varying FIR filter for the linearized perturbation model and the nominal state is directly obtained from the nonlinear nominal trajectory model. Since the FIR structured estimators use the finite horizon information on the most recent time interval, the proposed extended FIR filter satisfies the bounded input/bounded output (BIBO) stability, which can't be obtained from infinite impulse response (IIR) estimators. Thus, it can be expected that the proposed extended FIR filter is more robust than IIR structured estimators such as an extended Kalman filter for the round-of errors and the uncertainties from unknown initial states and uncertain system model parameters. The simulation results show that the proposed filter has better performance than the extended Kalman filter (EKF) in both robustness and fast convergency.

• 전력전자학회:학술대회논문집
• /
• 전력전자학회 1999년도 전력전자학술대회 논문집
• /
• pp.544-548
• /
• 1999

Using Observer, on the sensorless vector control system is a novel techniques for modern induction motor control. In this paper, a speed estimation algorithm of an induction motor using an extended kalman filter was proposed. Extended kalman filter can solve the problem, that have steady state error of estimated speed in flux and slip estimation method. The extended Kalman filter is employed to identify the speed of an induction motor and rotor flux based on the measured quantities such as stator current and DC link voltage. In order to confirming above proposal, computer simulation carried out using Matlab Simulink and show the effectiveness of the control drives for induction motor speed estimation.