• IEMEK Journal of Embedded Systems and Applications
• /
• v.6 no.1
• /
• pp.16-24
• /
• 2011

Automotive Radar Systems are currently under development for various applications to increase accuracy and reliability. The target tracking is most important in single or multiple target environments for accuracy. The tracking algorithm provides smoothed and predicted data for target position and velocity(Doppler). To this end, the fixed gain filter(${\alpha}{\beta}$ filter, ${\alpha}{\beta}{\gamma}$ filter) and dynamic filter(Kalman filter, Singer-Kalman filter, etc) are commonly used. Gating is used to decide whether an observation is assigned to an existing track or new track. Gating algorithms are normally based on computing a statistical error distance between an observation and prediction. The data association takes the observation-to-track pairings that satisfied gating and determines which observation-to-track assignment will actually be made. For data association, NNPDA(Nearest Neighbor Probabilistic Data Association) algorithm is proposed. In this paper, we designed a target tracking system developed for an Automotive Radar System. We show the experimental results of the 77GHz FMCW radar sensor on the roads. Four tracking algorithms(${\alpha}{\beta}$ filter, ${\alpha}{\beta}{\gamma}$ filter, 2nd order Kalman filter, Singer-Kalman filter) have been compared and analyzed to evaluate the performance in test scenario.

• Proceedings of the IEEK Conference
• /
• 2009.05a
• /
• pp.210-212
• /
• 2009

In object tracking based on the particle filter algorithm controlling the proper distribution of the samples is essential to accurately track the target. If the samples are spread too wide compared to the target size, the tracking accuracy may degrade as some samples can be caught by background clutters that is similar to the target. On the other hands if the samples are spread too narrow, the particle filter may fail to track the abrupt motion of the target. To solve this problem we propose an improved particle filter that adopts "re-weighting" technique at the observe step. We estimate the distribution of the weights of the current samples by its mean and variance. Then the samples are re-weighted so that the appropriate distribution of the samples in proportional to the target scale is obtained at the next select step. The proposed tracking method can avoid convergence to local mean and improve the accuracy of the estimated target state.

• 제어로봇시스템학회:학술대회논문집
• /
• 1992.10a
• /
• pp.424-429
• /
• 1992

A practical adaptive tracking filter for a maneuvering target is proposed in this paper by combining a modified input estimation technique with pseudo-residuals and a decoupled tracking filter in line-of-sight Cartesian coordinate system. Since the adaptive tracking filter has decoupled structure and computes maneuver input estimates for each axis separately, it requires much less computations compared with the coventional tracking filter with MIE technique without degrading performance. Also, since pseudo-measurement noises in line-of-sight Cartesian coordinate system are much less correlated compared with those of inertial Cartesian coordinate system, the proposed tracking filter produces less false alarms or miss detections to improve the performance.

• Journal of the Korea Institute of Military Science and Technology
• /
• v.14 no.4
• /
• pp.676-683
• /
• 2011

In this paper, we propose a robust method which can detect and track a high-speed small approaching target in a cluttered environment for Korean Active Protection System. The proposed method uses a temporal and spatial filter, tracking filter to detect and track a single target in consecutive order. And it is comprised of a candidate target detection step, a prior target selection step and a target tracking. Field tests on real infrared image sequences show that the proposed method could stably track a high speed and small target in complex background and target occlusion.

• 제어로봇시스템학회:학술대회논문집
• /
• 2000.10a
• /
• pp.99-99
• /
• 2000

An effect ive filter algorithm that can manage radar beam pointing efficiently is needed to track multi-target in the air. For effective beam management the filter has lobe good enough to predict future position of target and based on this filter output radar beam is control led to point toward the predicted target position in the air. In this paper, we investigate the ${\alpha}$-${\beta}$ filter known for its brief filter structure with the steady-state Kalman filter gain, the ruv filter, and the coordinate-transformed filter that can decouple the measurement noise variance.

• 제어로봇시스템학회:학술대회논문집
• /
• 2000.10a
• /
• pp.188-188
• /
• 2000

The most important problem in target tracking can be said to be modeling the tracking system correctly. Although the simple linear dynamic equation for this model has used until now, the satisfactory performance could not be obtained owing to uncertainties of the real systems in the case of designing the filters baged on the dynamic equations. In this paper, we propose the extended robust Kalman filter (ERKF) which can be applied to the real target tracking system with the parameter uncertainties. A nonlinear dynamic equation with parameter uncertainties is used to express the uncertain system model mathematically, and a measurement equation is represented by a nonlinear equation to show data from the radar in a Cartesian coordinate frame. To solve the robust nonlinear filtering problem, we derive the extended robust Kalman filter equation using the Krein space approach and sum quadratic constraint. We show the proposed filter has better performance than the existing extended Kalman filter (EKF) via 3-dimensional target tracking example.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.8 no.5
• /
• pp.1711-1725
• /
• 2014

Multi-target tracking is the main purpose of many video surveillance applications. Recently, multi-target tracking based on the particle filter method has achieved robust results by using the data association process. However, this method requires many calculations and it is inadequate for real time applications, because the number of associations exponentially increases with the number of measurements and targets. In this paper, to reduce the computational cost of the data association process, we propose a novel multi-target tracking method that excludes particle samples in the overlapped predictive region between the target to track and marginal targets. Moreover, to resolve the occlusion problem, we define an occlusion mode with the normal dynamic mode. When the targets are occluded, the mode is switched to the occlusion mode and the samples are propagated by Gaussian noise without the sampling process of the particle filter. Experimental results demonstrate the robustness of the proposed multi-target tracking method even in occlusion.

• 제어로봇시스템학회:학술대회논문집
• /
• 1991.10a
• /
• pp.125-130
• /
• 1991

This paper presents a tracking filter using pseudomeasurements in an estimated line-of-sight Cartesian coordinate system(ELCCS) whose x-axis is on the line-of-sight to an estimated target position. A target dynamics model and a measurement equation in the ELCCS are derived first and then a tracking filter in the ELCCS named moving coordinate tracking filter(MCTF) is proposed. It is shown that this MCTF is equivalent to a Kalman filter in the inertial Cartesian coordinate system which is widely used in the target tracking system. By approximating the MCTF for a pseudomeasurement noise and an error covariance matrix in the ELCCS, decoupling of three axes can be achieved. In this case, named decoupled moving coordinate tracking filter(DMCTF), computation time can be drastically reduced by utilizing its parallel structure. Finally, the stochastic properties of the MCTF and DMCTF are presented. Especially, a sufficient condition of nondestabilizing deviation for the DMCTF is proposed. The performance of the MCTF and DMCTF are compared with a conventional Kalman tracking filter.

• 제어로봇시스템학회:학술대회논문집
• /
• 1996.10b
• /
• pp.759-762
• /
• 1996

In previous work Kwon and Yoo [5] have shown that the FIR tracking algorithm using the input estimation technique. However, it has not solved the problem of systems with parameter uncertainties. Therefore, in this paper we propose a new robust $H_{\infty}$ FIR tracking filter to solve the target tracking problems under systems with parameter uncertainties. Also, we use here the input estimation approach to account for the possibility of maneuver. Simulation results show that the robust $H_{\infty}$ FIR tracking filter proposed here still has good tracking performance for a maneuvering target tracking problem even under all system parameter uncertainties.

• 제어로봇시스템학회:학술대회논문집
• /
• 2003.10a
• /
• pp.2340-2344
• /
• 2003

This paper considers a tracking filter algorithm which can track a maneuvering target. Multiresolutional Interacting Multiple Model (MRIMM) algorithm is proposed to reduce computational burden. In this paper multiresolutional state space model equation and multiresolutional measurement equation are derived by using wavelet transform. This paper shows the outline of MRIMM algorithm. Simulation results show that MRIMM algorithm maintains a good tracking performance and reduces computational burden.