A dual channel perturbation particle filter algorithm based on GPU acceleration

doi:10.21629/JSEE.2018.04.20

Journal of Systems Engineering and Electronics ›› 2018, Vol. 29 ›› Issue (4): 854-863.doi: 10.21629/JSEE.2018.04.20

• Software Algorithm and Simulation • Previous Articles Next Articles

A dual channel perturbation particle filter algorithm based on GPU acceleration

Fan LI^1,*(), Hongkui BI²(), Jiajun XIONG²(), Chenlong YU¹(), Xuhui LAN²()

¹ Department of Graduate, Air Force Early-warning College, Wuhan 430019, China
² Department of Early Warning Intelligence, Air Force Early-warning College, Wuhan 430019, China

Received:2017-03-30 Online:2018-08-01 Published:2018-08-30
Contact: Fan LI E-mail:1746338543@qq.com;bhk001@126.com;13871163420@139.com;342583844@qq.com;Lansoft007@sohu.com
About author:LI Fan was born in 1992. He received his B.S degree and M.S degree from the Air Force Earlywarning College in 2010 and 2016, respectively. Now, he is a Ph.D. candidate and a lecturer in Department of Early Warning Intelligence at the Air Force Early-warning College. His research interest is the near space hypersonic vehicle early warning detection technology. E-mail: 1746338543@qq.com|BI Hongkui was born in 1964. She received her B.S and M.S degrees from Huazhong University of Science and Technology in 1986 and 1989, respectively. Now, she is a professor in Air Force Early-warning College. Her research interests include radar data processing, radar equipment technology and application. E-mail: bhk001@126.com|XIONG Jiajun was born in 1961. He received his B.S. degree in computer software from Huazhong University of Science and Technology in 1883, M.S degree from National University of Defense Technology in 1988, and Ph.D degree from Huazhong University of Science and Technology in 2004. Now, he is a professor in Air Force Early-warning College. His research interests include data fusion and early warning intelligence analysis. E-mail: 13871163420@139.com|YU Chenlong was born in 1989. He received his B.S. degree from South China University of Technology in 2012 and M.S. degree from Air Force Early-warning College in 2015. Now he is a Ph.D. candidate in Air Force Early-warning College. His research interests include signal processing, target tracking and target detection. E-mail: 342583844@qq.com|LAN Xuhui was born in 1976. He received his B.S degree, M.S degree and Ph.D. degree from Air Force Early-warning College in 1999, 2002, and 2012 respectively. Now he is an associate professor in Air Force Early-warning College. His research directions are multi-sensor data fusion and integrated electronic information system. E-mail: Lansoft007@sohu.com
Supported by:
the National High-tech R & D Program of China(2015AA7056045);the National High-tech R & D Program of China(2015AA8017032P);the National Natural Science Foundation of China(61401504);This work was supported by the National High-tech R & D Program of China (2015AA7056045; 2015AA8017032P), and the National Natural Science Foundation of China (61401504)

Abstract

Abstract:

The particle filter (PF) algorithm is one of the most commonly used algorithms for maneuvering target tracking. The traditional PF maps from multi-dimensional information to onedimensional information during particle weight calculation, and the incorrect transmission of information leads to the fact that the particle prediction information does not match the weight information, and its essence is the reduction of the information entropy of the useful information. To solve this problem, a dual channel independent filtering method is proposed based on the idea of equalization mapping. Firstly, the particle prediction performance is described by particle manipulations of different dimensions, and the accuracy of particle prediction is improved. The improvement of particle degradation of this algorithm is analyzed in the aspects of particle weight and effective particle number. Secondly, according to the problem of lack of particle samples, the new particles are generated based on the filtering results, and the particle diversity is increased. Finally, the introduction of the graphics processing unit (GPU) parallel computing the platform, the "channel-level" and "particlelevel" parallel computing the program are designed to accelerate the algorithm. The simulation results show that the algorithm has the advantages of better filtering precision, higher particle efficiency and faster calculation speed compared with the traditional algorithm of the CPU platform.

Key words: particle filter (PF), dual channel filtering, graphic processing unit (GPU), parallel operation

Fan LI, Hongkui BI, Jiajun XIONG, Chenlong YU, Xuhui LAN. A dual channel perturbation particle filter algorithm based on GPU acceleration[J]. Journal of Systems Engineering and Electronics, 2018, 29(4): 854-863.

Figures/Tables 11

Fig 1

Fig 2

Fig 3

Fig 4

Fig 5

Fig 6

Fig 7

Fig 8

Fig 9

Table 1

Fig 10

References 32

1	CAPPE O, GODSILL S J, MOULINES E. An overview of existing methods and recent advances in sequential Monte Carlo. Proceedings of the IEEE, 2007, 95 (5): 899- 924. doi: 10.1109/JPROC.2007.893250
2	DOUCET A, JOHANSEN A M. A tutorial on particle filtering and smoothing: 15 years later. Handbook of Nonlinear Filtering, 2011, 1- 39.
3	GUSTAFSSON F. Particle filter theory and practice with positioning applications. IEEE Aerospace and Electronic Systems Magazine, 2010, 25 (7): 53- 81. doi: 10.1109/MAES.2010.5546308
4	LI T C, BOLIC M, DJURIC P M. Resampling methods for particle filtering: classification, implementation, and strategies. IEEE Signal Processing Magazine, 2015, 32 (3): 70- 86. doi: 10.1109/MSP.2014.2330626
5	LI T C, VILLARRUBIA G, SUN S D, et al. Resampling methods for particle filtering: identical distribution, a new method, and comparable study. Frontiers of Information Technology and Electronic Engineering, 2015, 16 (11): 969- 984. doi: 10.1631/FITEE.1500199
6	LI T C, SUN S D, SATTAR T P, et al. Fight sample degeneracy and impoverishment in particle filters: a review of intelligent approaches. Expert Systems with Applications, 2014, 41 (8): 3944- 3954. doi: 10.1016/j.eswa.2013.12.031
7	LI X R, JILKOV V P. A survey of maneuvering target tracking, part VIc: approximate nonlinear density filtering in discrete time. Proc. of the SPIE Defense, Security, and Sensing, 2012: 1-12.
8	SARKKA S. On unscented Kalman filtering for state estimation of continuous-time nonlinear systems. IEEE Trans. on Automatic Control, 2007, 52(9): 1631-1641.
9	BARRETT L, HENZI S P. Particle filters for positioning, navigation and tracking. IEEE Trans. on Signal Processing, 2002, 50(2): 425-437.
10	ARASARATNAM I, HAYKIN S. Cubature Kalman filters. IEEE Trans. on Automatic Control, 2009, 54(6): 1254-1269.
11	ZHANG W J, WANG S Y, FENG Y L, et al. Huber-based high-degree cubature Kalman tracking algorithm. Acta Physica Sinica, 2016, 65 (8): 088401.
12	XU C, GAO M, YANG Y. Self-turning hierarchical Kalmanparticle for efficient target tracking. Infrared and Laser Engineering, 2015, 44 (6): 1942- 1949.
13	ZUO J Y, JIA Y N. Particle filter guided by iterated extended Kalman filter. Proc. of the 13th Automation and Systems International Conference on Control, 2013: 1605-1609.
14	LIU C Y, SHUI P L, LI S. Unscented extended Kalman filter for target tracking. Journal of Systems Engineering and Electronics, 2011, 22 (2): 188- 192.
15	LI T C, SUN S D. Double-resampling based Monte Carlo localization for mobile robot. Acta Automatma Sinica, 2010, 36 (9): 1279- 1286.
16	LI T C, FAN H Q, SUN S D. Particle filtering: theory, approach, and application for multi-target tracking. Acta Automatma Sinica, 2015, 41 (12): 1981- 2002.
17	JILKOV V P, WU J D. Efficient GPU-accelerated implementation of particle and particle flow filters for target tracking. Journal of Advances in Information Fusion, 2015, 10 (1): 73- 88.
18	BROWN J A, CAPSON D W. A framework for 3D modelbased visual tracking using a GPU-accelerated particle filter. IEEE Trans. on Visualization & Computer Graphics, 2011, 18(1): 68-80.
19	BASHI A S, JILKOV V P, LI X R, et al. Distributed implementations of particle filters. Proc. of the International Conference on Information Fusion, 2003: 1164-1171.
20	DEL M P, DOUCET A. Particle methods: an introduction with applications. Esaim: Proceedings, 2014, 44, 1- 46.
21	ZHANG Q, QIAO Y K, KONG X Y, et al. Study on stochastic perturbation strong tracking particle filter. Acta Physica Sinica, 2014, 63 (11): 110505. doi: 10.3390/app7111152
22	BOLIC M, DJURIC P M, HONG S. Resampling algorithms and architectures for distributed particle filters. IEEE Trans. on Signal Processing, 2005, 53(7): 2442-2450.
23	STEINBRING J, HANEBECK U D. GPU-accelerated progressive Gaussian filtering with applications to extended object tracking. Proc. of the IEEE International Conference on Information Fusion, 2015: 1038-1045.
24	TAGER S. GPU-specfic Kalman filtering and retrodiction for large-scale target tracking. Electrical & Computer Engineering, 2013, 1- 113.
25	HASHMI M F, PAL R, SAXENA R, et al. A new approach for real time object detection and tracking on high resolution and multi-camera surveillance videos using GPU. Journal of Central South University, 2016, 23 (1): 130- 144.
26	GOYAL B, BUDHRAJA T, BHATNAGAR R, et al. Implementation of particle filters for single target tracking using CUDA. Proc. of the 5th IEEE International Conference on Advances in Computing and Communications, 2016: 28-32.
27	KIM S, CHO J, PARK D. Moving-target position estimation using GPU-based particle filter for IOT sensing applications. Applied Sciences, 2017, 7 (11): 1- 14. doi: 10.3390/app7111152
28	LIU K, WEI S, CHEN Z, et al. A real-time high performance computation architecture for multiple moving target tracking based on wide-area motion imagery via cloud and graphic processing units. Sensors, 2017, 17 (2): 1- 16. doi: 10.3390/s17020356
29	TAN F, XIAO S, LI L. Enhanced CMT tracking algorithm with CUDA acceleration. Proc. of the IEEE International Conference on Multimedia & Expo Workshops, 2017: 441-446.
30	COOK S. CUDA programming: a developer's guide to parallel computing with GPUs. San Francisco: Morgan Kaufmann, 2012.
31	ZOU W J, GONG X, BO Y M. Adaptive layered-sampling auxiliary particle filter's research and application in video tracking. Acta Photonica Sinica, 2010, 39 (3): 571- 576. doi: 10.3788/gzxb20103903.0571
32	LI F, BI H K, DUAN M, et al. Particle filter algorithm based on sequential quadratic programming. Modern Radar, 2016, 38 (9): 50- 56.

Number of particles		1 000	500	100
Parameter	Algorithm	1 000	500	100
$X$_$_{\rm ORMSE}$	SIR	346.53	393.45	—
	APF	327.45	374.81	—
	SQP-PF	309.21	352.26	—
	GPU-DC-PF	263.64	321.63	357.21
$Y$_$_{\rm ORMSE}$	SIR	351.64	402.33	—
	APF	336.10	389.02	—
	SQP-PF	317.07	365.31	—
	GPU-DC-PF	283.51	335.21	361.84
$N_{e}$	SIR	189	84	—
	APF	257	112	—
	SQP-PF	367	137	—
	GPU-DC-PF	583	257	68
$f$	SIR	39	69	100
	APF	23	54	100
	SQP-PF	12	37	100
	GPU-DC-PF	0	0	23
$T_{c}$	SIR	50.814	27.364	—
	APF	58.312	29.507	—
	SQP-PF	60.089	30.012	—
	GPU-DC-PF	12.731	8.542	2.688

[1]	Ting SU, Yong GAO. TDOA estimation of dual-satellites interference localization based on blind separation [J]. Journal of Systems Engineering and Electronics, 2019, 30(4): 696-702.
[2]	Wei Han, Ziyue Tang, and Zhenbo Zhu. Method of target tracking with Doppler blind zone constraint [J]. Journal of Systems Engineering and Electronics, 2013, 24(6): 889-898.

A dual channel perturbation particle filter algorithm based on GPU acceleration

RichHTML

PDF (PC)

Knowledge

Abstract

Cite this article

Share this article

Figures/Tables 11

References 32

Related Articles 2

Recommended Articles

Metrics

Comments