Performance Comparison of Various Extended Kalman Filter and Cost-Reference Particle Filter for Target Tracking with Unknown Noise

노이즈 불확실성하에서의 확장칼만필터의 변종들과 코스트 레퍼런스 파티클필터를 이용한 표적추적 성능비교

  • Received : 2018.04.18
  • Accepted : 2018.08.02
  • Published : 2018.09.30


In this paper, we study target tracking in two dimensional space using a Extended Kalman filter(EKF), various Extended Kalman Filter and Cost-Reference Particle Filter(CRPF), which can effectively estimate the state values of nonlinear measurement equation. We introduce various Extended Kalman Filter which the Unscented Kalman Filter(UKF), the Central Difference Kalman Filter(CDKF), the Square Root Unscented Kalman Filter(SR-UKF), and the Central Difference Kalman Filter(SR-CDKF). In this study, we calculate Mean Square Error(MSE) of each filters using Monte-Carlo simulation with unknown noise statistics. Simulation results show that among the various of Extended Kalman filter, Square Root Central Difference Kalman Filter has the best results in terms of speed and performance. And, the Cost-Reference Particle Filter has an advantageous feature that it does not need to know the noise distribution differently from Extended Kalman Filter, and the simulation result shows that the excellent in term of processing speed and accuracy.

본 논문에서는 비선형성을 가지는 측정방정식의 상태값을 효과적으로 추정할 수 있는 확장칼만필터(Extended Kalman Filter/EKF)와 확장칼만필터의 변종들 그리고 코스트 레퍼런스 파티클필터(Cost-Reference Particle Filter/CRPF)를 이용하여 이차원 공간에서 표적추적 성능에 관하여 연구한다. 확장칼만필터의 변종으로 분산점칼만필터(Unscented Kalman Filter/UKF), 중심차분칼만필터(Central Difference Kalman Filter/CDKF), 제곱근 분산점칼만필터(Square Root Unscented Kalman Filter/SR-UKF) 그리고 제곱근 중심차분칼만필터(Square Root Central Difference Kalman Filter/SR-CDKF)를 소개한다. 본 연구에서는 노이즈가 불확실한 표적에 대하여 몬테카를로 시뮬레이션 기법을 이용하여 각 필터들의 평균제곱오차(Mean Square Error/MSE)를 계산하였다. 시뮬레이션 결과 확장칼만필터의 변종들 중에서 제곱근 중심차분칼만필터가 속도와 성능 면에서 가장 우수한 결과를 보여주었다. 코스트 레퍼런스 파티클 필터는 확장칼만필터와 다르게 노이즈의 확률 분포를 알 필요가 없다는 유리한 특성을 가지고 있으며 시뮬레이션 결과 제곱근 중심차분칼만필터보다 처리속도 및 정확도 면에서 우수한 결과를 보여주었다.



Supported by : 해양연구소


  1. K.Ito and K. Xiong (2000) "Gaussian filters for nonlinear filterinf problem", IEEE Transactions on Automatic Control, 45(5), 910-927
  2. S.J. Julier, J. K. Uhlmann and H. F. Durrant-Whyte (1995) "A new approach for filtering nonlinear systems", Proceedings of the American Control Conference, 3, 1628-1632
  3. R. E. Kalman (1960) "A new approach to linear filtering and prediction", Transactions of the ASME-Journal of Basic Engineering, 82, 35-45
  4. J. Lim (2013) "A target tracking base on bearing and range measurement with unknown noise statistic", Journal of Electrical Engineering and Technology, 8(6), 1520-1529
  5. J. Lim, M. Shin and W. Hwang (2017) "Variants of extended Kalman filtering approaches for Bayesian tracking", International Journal of Robust and nonlinear Control, 27(2), 319-346
  6. R. Van der Merwe and E. A. Wan (2001) "The Squareroot unscented Kalman fitler for state and parameterestimation". IEEE International Conference on Acoustics, Speech and Signal Processing, 6, 3461-3464
  7. R. Van der Merwe (2004) "Sigma-point Kalman filters for probabilistic inference in dynamic state-space medels", Science & Engineering, Oregon health & Science University, Porland
  8. M. Norgaard, N. Poulsen and O. Rayn (2000) "Nwe developments in state estimation for nonlinear systems", Automatica, 36(11)
  9. S. Xu, (2006) "Particle filtering for systems with unknown noise probability distribution", Electrical and Computer Engineering, Stony Brook University