Applied Technology in an Adaptive Particle Filter Based on Interval Estimation and KLD-Resampling

Abstract:

Particle filter as a sequential Monte Carlo method is widely applied in stochastic sampling for state estimation in a recursive Bayesian filtering framework. The efficiency and accuracy of the particle filter depends on the number of particles and the relocating method. The automatic selection of sample size for a given task is therefore essential for reducing unnecessary computation and for optimal performance, especially when the posterior distribution greatly varies overtime. This paper presents an adaptive resampling method (IE_KLD_PF) based on interval estimation, and after interval estimating the expectation of the system states, the new algorithm adopts Kullback-Leibler distance (KLD) to determine the number of particles to resample from the interval and update the filter results by current observation information. Simulations are performed to show that the proposed filter can reduce the average number of samples significantly compared to the fixed sample size particle filter.

• Periodical: Advanced Materials Research (Volume 1014)
• Pages: 452-458
• DOI: https://doi.org/10.4028/www.scientific.net/AMR.1014.452
• Online since: July 2014
• Authors: Jian Zhang*, Ling Shen
• Keywords: Interval Estimation, KLD, Particle Filter, Resampling Strategy

References

[1] H. Dandach, F. Abdallah, J. D. Miras, A. Charara, in: Vehicle Dynamics Estimation Using Box Particle Filter. 2012 12th International Conference on Control Automation Robotics & Vision (ICARCV). Guangzhou, China. (2012).

DOI: 10.1109/icarcv.2012.6485144

[2] Olivier Cappe, Simon J. Godsill, Eric Moulines. An Overview of Existing Methods and Recent Advances in Sequential Monte Carlo. Proceedings of the IEEE. Vol. 95(2007), pp.899-924.

DOI: 10.1109/jproc.2007.893250

[3] F. Abdallah, A. Gning, P. Bonnifait. Box particle filtering for nonlinear state estimation using interval analysis. Automatica. Vol. 44(2008), p.807–815.

DOI: 10.1016/j.automatica.2007.07.024

[4] D. Fox. Adapting the sample size in particle filters through KLD-sampling. Int. J. Robot. Res. Vol. 22(2003), pp.985-1003.

DOI: 10.1177/0278364903022012001

[5] L. Jing, P. Vadakkepat. Interacting MCMC particle filter for tracking maneuvering target. Digital Signal Process. Vol. 20(2010), p.561–574.

DOI: 10.1016/j.dsp.2009.08.011

[6] T. Li, T.P. Sattar, S. Sun. Deterministic resampling: unbiased sampling to avoid sample impoverishment in particle filters. Signal Process. Vol. 92(2012), p.1637–1645.

DOI: 10.1016/j.sigpro.2011.12.019

[7] M.K. Pitt, N. Shephard. Filtering via simulation: auxiliary particle filters. J. Am. Stat. Assoc. Vol. 94(1999), p.590–599.

DOI: 10.1080/01621459.1999.10474153

[8] M.S. Arulampalam, S. Maskell, N. Gordon, T. Clapp. A tutorial on particle filters for online nonlinear/non-Gaussian Bayesian tracking. IEEE Trans. Signal Process. Vol. 50(2002), p.174–188.

DOI: 10.1109/78.978374

[9] M.S. Arulampalam, S. Maskell, N. Gordon, T. Clapp. A Tutorial on Particle Filters for Online Nolinear/Non-Gaussian Bayesian Tracking. IEEE Transactions on Signal Processing. February, Vol. 50(2002), pp.174-188.

DOI: 10.1109/78.978374

[10] M. Pitt, N. Shephard. Filtering via simulation: Auxiliary particle filters. J. Amer. Statist. Assoc., Vol. 94(1999), pp.590-599.

DOI: 10.1080/01621459.1999.10474153

[11] C. Kwok, D. Fox, M. Meila. Adaptive real-time particle filters for robot localization. IEEE Int. Conf. on Robotics and Automation, Taipei, Taiwan. Vol. 2(2003), p.2836–2841.

DOI: 10.1109/robot.2003.1242022