• Smart Media Journal
• /
• v.13 no.5
• /
• pp.9-18
• /
• 2024

Multi-Object trajectory modeling is a major challenge in MOT. CenterTrack tried to solve this problem with a Heatmap-based method that tracks the object center position. However, it showed limited performance when tracking objects with complex movements and nonlinearities. Considering the degradation factor of CenterTrack as the dynamic movement of pedestrians, we integrated the EKF into CenterTrack. To demonstrate the superiority of our proposed method, we applied the existing KF and UKF to CenterTrack and compared and evaluated it on various datasets. The experimental results confirmed that when EKF was integrated into CenterTrack, it achieved 73.7% MOTA, making it the most suitable filter for CenterTrack.

• Journal of the Institute of Electronics Engineers of Korea SC
• /
• v.46 no.3
• /
• pp.52-58
• /
• 2009

If a moving target has a linear characteristics, the Kalman filter can estimate relatively accurate the location of a target, but this performance depends on how the dynamic status characteristics of the target is accurately modeled. In many practical problems of tracking a maneuvering target, a simple kinematic model can fairly accurately describe the target dynamics for a wide class of maneuvers. However, since the target can exhibit a wide range of dynamic characteristics, no fixed SKF(Simple Kalman filter) can be matched to estimate, to the required accuracy, the states of the target for every specific maneuver. In this paper, a new AKF(Active Kalman filter) is proposed to solve this problem The process noise covariance level of the Kalman filter is adjusted at each time step according to the study result which uses the neural network algorithm. It is demonstrated by means of a computer simulation that the tracking capability of the proposed AKF(Active Kalman filter) is better than that of the SKF(Simple Kalman Filter).

• Proceedings of the Korean Institute of Navigation and Port Research Conference
• /
• 2016.05a
• /
• pp.302-304
• /
• 2016

The purpose of this paper is to draw comparison between the performance of ${\alpha}-{\beta}-{\gamma}$ filter and Kalman filter of a tracking module for ARPA system on board high dynamic warship. The comparison is based on the filters' capability to reduce residual error and maintain a stable transient response. The residual error is computed from the difference between the observed the predicted positions for the entire tracking period. The results indicate that the Kalman filter has a higher tracking accuracy compared to the optimal ${\alpha}-{\beta}-{\gamma}$ filter. However, both filters have a similar transient response.

• Journal of Astronomy and Space Sciences
• /
• v.25 no.3
• /
• pp.283-290
• /
• 2008

Satellite formation flying can provide the platform for interferometric observation to acquire the precise data and ensure the flexibility for space mission. This paper presents development and verification of an algorithm to estimate the baseline between formation flying satellites. To estimate a baseline(relative navigation) in real time, EKF(Extended Kalman Filter) and UKF(Unscented Kalman Filter) are used. Measurements for updating a state-vector in Kalman Filter are GPS single difference data. In results, The position errors in estimated baseline are converged to less than ${\pm}1m$ in both EKF and UKF. And as using the two types of Kalman filter, it is clear that the unscented Kalman filter shows a relatively better performance than the extended Kalman filter by comparing an efficiency to the model which has a non-linearity.

• Journal of Astronomy and Space Sciences
• /
• v.22 no.4
• /
• pp.501-512
• /
• 2005

This research supposed when a fictitious KSIV-I space launch vehicle launches from NARO space center. This compared and analyzed the results from real-time trajectory estimation using the Extended Kalman Filter and the Unscented Kalman Filter. A virtual trajectory and observation data are generated for the fictitious KSLV-I and three measurement radars. The performances of both Otters are compared for several simulations with small initial errors, large initial errors, 20Hz and 10Hz data rate. The results show that the Unscented Kalman Filter yields faster convergence and more accurate than the Extended Kalman Filter for the cases with larger initial error and slower data rate conditions.

• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.40 no.2
• /
• pp.108-117
• /
• 2012

This paper focuses on a performance analysis of TRN among various nonlinear filtering methods. In a TRN research, extended Kalman filter(EKF) is a basic estimation algorithm. In this paper, iterated EKF(IEKF), EKF with stochastic linearization(SL), and unscented Kalman filter(UKF) algorithms are introduced to compare navigation performance with original EKF. In addition to introduced sequential filters, bank of Kalman filters method, which is one of the batch method, is also presented. Finally, by simulating an artificial aircraft mission, EKF with SL was chosen as the most consistent filter in the introduced sequential filters. Also, results suggested that the bank of Kalman filters can be alternative for TRN, when a fast convergence of navigation solution is needed.

• 제어로봇시스템학회:학술대회논문집
• /
• 1997.10a
• /
• pp.1679-1683
• /
• 1997

In this paper a linear Kalman filter for calibration of gimballed inertial navigation system(GINS) is designed and its performace is analyzed through the simulation with a real navigation data. Simulation results show that the proposed Kalman filter gives a good performance to calibrate the sensor errors.

• 제어로봇시스템학회:학술대회논문집
• /
• 1988.10a
• /
• pp.187-191
• /
• 1988

To track the manenvering target and to derive the Filter using state estimation and information in real time, we derive adaptive Kalman Filter which reinitialize the internal filter mode.

• Journal of Institute of Control, Robotics and Systems
• /
• v.9 no.11
• /
• pp.917-924
• /
• 2003

Recurrent neural networks have been successfully applied to communications channel equalization. Major disadvantages of gradient-based learning algorithms commonly employed to train recurrent neural networks are slow convergence rates and long training sequences required for satisfactory performance. In a high-speed communications system, fast convergence speed and short training symbols are essential. We propose decision feedback equalizers using a recurrent neural network trained with Kalman filtering algorithms. The main features of the proposed recurrent neural equalizers, utilizing extended Kalman filter (EKF) and unscented Kalman filter (UKF), are fast convergence rates and good performance using relatively short training symbols. Experimental results for two time-varying channels are presented to evaluate the performance of the proposed approaches over a conventional recurrent neural equalizer.

• Journal of Institute of Control, Robotics and Systems
• /
• v.12 no.3
• /
• pp.201-207
• /
• 2006

A robust Kalman filtering method for uncertain stochastic systems is suggested by adopting a perturbation estimation process which is to reconstruct total uncertainty with respect to the nominal state transition equation. The predictor and corrector of discrete Kalman filter are reformulated with the perturbation estimator. Successively, the state and perturbation estimation error dynamics and the corresponding error covariance propagation equations are derived as well. Finally we have the recursive algorithm of Combined Kalman Filter-Perturbation Estimator (CKF). The proposed combined Kalman filter-perturbation estimator has the property of integrating innovations and the adaptation capability to system uncertainties. A numerical example is shown to demonstrate the effectiveness of the proposed scheme.