• Journal of the Korean Institute of Intelligent Systems
• /
• v.15 no.5
• /
• pp.552-559
• /
• 2005

Recurrent neural networks(RNNs) trained with gradient based such as real time recurrent learning(RTRL) has a drawback of slor convergence rate. This algorithm also needs the derivative calculation which is not trivialized in error back propagation process. In this paper a derivative free Kalman filter, so called the unscented Kalman filter(UKF), for training a fully connected RNN is presented in a state space formulation of the system. A derivative free Kalman filler learning algorithm makes the RNN have fast convergence speed and good tracking performance without the derivative computation. Through experiments of nonlinear channel equalization, performance of the RNNs with a derivative free Kalman filter teaming algorithm is evaluated.

• Journal of international Conference on Electrical Machines and Systems
• /
• v.3 no.4
• /
• pp.458-464
• /
• 2014

This paper proposes the rotor position and rotor speed estimation for an interior permanent magnet synchronous machines (IPMSM) using Unscented Kalman Filter (UKF) in alpha-beta coordinate system. Conventional algorithms using UKF are based on the simple observer model of IPMSM in d-q coordinate system. Rotor acceleration is neglected within the sampling step. An expansion of the observer model in an alpha-beta coordinate system with the consideration of the rotor speed variation provides the improved rotor position and speed estimation. The results show good stability concerning the expansion of observer model for the IPMSM.

• Journal of the Korea Institute of Military Science and Technology
• /
• v.17 no.1
• /
• pp.96-107
• /
• 2014

Tracking the tactical object is a fundamental affair in network-equipped modern warfare. Geodetic coordinate system based on longitude, latitude, and height is suitable to represent the location of tactical objects considering multi platform data fusion. The motion of tactical object described as a dynamic model requires an appropriate filtering to overcome the system and measurement noise in acquiring information from multiple sensors. This paper introduces the filter suitable for multi-sensor data fusion and tactical object tracking, particularly the unscented transform(UT) and its detail. The UT in Unscented Kalman Filter(UKF) uses a few samples to estimate nonlinear-propagated statistic parameters, and UT has better performance and complexity than the conventional linearization method. We show the effects of UT-based filtering via simulation considering practical tactical object tracking scenario.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.17 no.6
• /
• pp.1273-1280
• /
• 2013

The UKF algorithm for tracking moving objects has fast convergence speed and good tracking performance without the derivative computation. However, this algorithm has serious drawbacks which limit its use in conditions such as Gaussian noise distribution. Meanwhile, the particle filter(PF) is a state estimation method applied to nonlinear and non-Gaussian systems without these limitations. But this method also has some disadvantages such as computation increase as the number of particles rises. In this paper, we propose the Unscented Particle Filter (UPF) algorithm which combines Unscented Kalman Filter (UKF) and Particle Filter (PF) in order to overcome these drawbacks.The performance of the UPF algorithm was tested to compare with Particle Filter using a 2-DOF (Degree of Freedom) Pendulum System. The results show that the proposed algorithm is more suitable to the nonlinear and non-Gaussian state estimation compared with PF.

• Journal of the Institute of Electronics Engineers of Korea SC
• /
• v.48 no.6
• /
• pp.82-90
• /
• 2011

This paper proposes an advanced outdoor localization algorithm of a GPS(global positioning system)-INS(inertial navigation system) integration system. In order to reduce noise from the internal INS sensors, discrete wavelet transform and variable threshold method are utilized. The UPF (unscented particle filter) combines GPS information and INS signals to implement precise outdoor localization algorithm and to reduce noise caused by the acceleration, deceleration, and unexpected slips. The conventional de-noising method is mainly carried out using a low pass filter and a high pass filter which essentially result in signal distortions. This newly proposed system utilizes the vibration information of actuator according to fluctuations of the velocity to minimize signal distortions. The UPF also resolves non-linearities of the actuator and non-normal distributions of noises. Effectiveness of the proposed algorithm has been verified through the real experiments and the results are demonstrated.

• Journal of the Korea Society for Simulation
• /
• v.19 no.4
• /
• pp.199-208
• /
• 2010

Positions of tactical objects are represented as Time, Space and Position Information(TSPI) in modeling and simulations(M&S). The format and required information record for TSPI is investigated by referring the TSPI object model of the Test and Training Enabling Architecture(TENA), which has been developed by the United States Department of Defense. The most sophisticated tactical data link, Link-16 has a Precise Participant Location and Information (PPLI) message. We study the data format for exchanging TSPI data based on the PPLI message. To evaluate and track positions of tactical objects, we consider the Kalman filter for linear systems, and the extended Kalman filter and the unscented Kalman filter for nonlinear systems. Based on motion equations of a ballistic missile, the tracking performance for the trajectory of the ballistic missile is simulated by the unscented Kalman filter.

• Proceedings of the Korean Society of Precision Engineering Conference
• /
• 2004.05a
• /
• pp.166-166
• /
• 2004

운전자 보조시스템에는 적응순항제어 (adaptive cruise control), 차선변경 (lane change), 충돌경고 (collision warning), 충돌회피 (collision avoidance), 및 자동주차 (automatic parking) 등이 있다. 이런 운전자 보조시스템은 어떤 목적을 가지고 있다. 운전자의 부담을 줄이고 안전을 위하여 차량의 주행방향에 있는 장애물이나 차량을 감지하여 차량간의 안전거리론 유지하고 자동차가 일정 속도를 유지하도록 한다. 운전자 보조시스템의 효율은 센서들로부터 얻어진 정보의 해석에 달려있다.(중략)

• The Journal of Korea Robotics Society
• /
• v.9 no.4
• /
• pp.216-224
• /
• 2014

This paper compares methods for attitude estimation of a UUV(Unmanned Underwater Vehicle). Attitude estimation plays a key role in underwater navigation using DVL(Doppler Velocity Log). The paper proposes attitude estimation methods using EKF(Extended Kalman Filter), UKF(Unscented Kalman Filter), and CF(Complementary Filter). It derives methods using the measurements from MEMS-AHRS(Microelectromechanical Systems-Attitude Heading Reference System) and DVL. The methods are used for navigation in a test pool and their navigation performance is compared. The results suggest that even if there is no measurement relative to some absolute landmarks, DVL-only navigation can be useful for navigation in a limited time and range.

• 제어로봇시스템학회:학술대회논문집
• /
• 2003.10a
• /
• pp.2331-2334
• /
• 2003

The basic difference between the EKF(Extended Kalman Filter) and UKF(Unscented Kalman Filter) stems from the manner in which Gaussian random variables(GRV) are represented for propagating through system dynamics. In the EKF, the state distribution is approximated by a GRV, which is then propagated analytically through the first-order linearization of the nonlinear system. This can possibly introduce large errors in the true posterior mean and covariance of the transformed GRV, which may lead to sub-optimal performance and sometimes divergence of the filter. However, the UKF addresses this problem by using a deterministic sampling approach. The state distribution is also approximated by a GRV, but is now represented using a minimal set of carefully chosen sample points. These sample points completely capture the true mean and covariance of the GRV, and UKF captures the posterior mean and covariance accurately up to the 2nd order(Taylor series expansion) for any nonlinearity. This paper utilizes the UKF to determine spacecraft orbit when only magnetometer is available. Several catastrophic failures of spacecraft in orbit have been attributed to failures of the spacecraft mission. Recently studies on contingency-major sensor failure cases- have been performed. For mission success, contingency design or plan should be implemented in case of a major sensor failure. Therefore the algorithm presented in this paper can be used for a spacecraft without GPS or contingency design in case of GPS failure.

• Journal of the Korean Society for Precision Engineering
• /
• v.29 no.10
• /
• pp.1128-1136
• /
• 2012

In this paper, the states and parameters in a dynamic system are estimated by applying an Unscented Kalman Filter (UKF). The UKF is widely used in various fields such as sensor fusion, trajectory estimation, and learning of Neural Network weights. These estimations are necessary and important in determining the stability of a mobile system, monitoring, and predictions. However, conventional approaches are difficult to estimate based on the experimental data, due to properties of non-linearity and measurement noises. Therefore, in this paper, UKF is applied in estimating the states and parameters needed. An experimental dynamic system has been set up for obtaining data and the experimental data is collected for parameter estimation. The measurement noises are primarily reduced by applying the Low Pass Filter (LPF). Given the simulation results, the estimated error rate is 39 percent more efficient than the results obtained using the Least Square Method (LSM). Secondly, the estimated parameters have an average convergence period of four seconds.