拟蒙特卡罗-高斯粒子滤波算法研究及其硬件实现
doi: 10.3724/SP.J.1146.2009.01002
Research and Hardware Implementation of Quasi-Monte-Carlo Gaussian Particle Filter
-
摘要: 该文针对粒子滤波计算量大，难以在工程中应用的问题，用拟蒙特卡罗采样(QMC)代替蒙特卡罗采样(MC)，减少了运算量。分析并给出了拟蒙特卡罗-高斯粒子滤波(QMC-GPF)算法的并行结构。在该并行结构的基础上，研究了基于FPGA的QMC-GPF的设计与实现。在实现过程中选取2作基数来产生Faure序列，将乘法运算、求模运算简化为便于在FPGA中实现的按位异或运算；采用查找表实现指数函数等复杂函数的计算，充分利用了FPGA中大量的Block RAM资源；给出了Cholesky分解矩阵各元素的并行计算结构。以红外图像弱小目标跟踪实验为例，验证了本设计的有效性和实时性。Abstract: A large amount of computation of particle filter limits its engineering application. According to this problem, Quasi-Monte Carlo (QMC) sampling is used to replace Monte Carlo (MC) sampling, reducing the required computation. Quasi-Monte-Carlo Gaussian Particle Filter (QMC-GPF) algorithm parallel architecture is proposed. Based on the parallel architecture, this paper lays emphasis on the implementation of the algorithm on FPGA in detail. Base 2 is used to generate Faure sequences, thus instead of multiplication and mod only bitwise XOR, which is easily to realize on FPGA, is needed to generate the sequences. Look-up tables are used in calculating the complex functions such as exponential function, which makes full use of the large number of Block RAM of FPGA. The parallel structure is designed to compute the elements of the Cholesky decomposition matrix. Infrared imaging dim small target tracking is realized on FPGA and the results show the efficiency and real time of the design.
-
李良群, 黄敬雄, 谢维信. 被动传感器阵列中基于粒子滤波的目标跟踪[J].电子与信息学报.2009, 31(4):844-847浏览[2]Yardim C, Gerstoft P, and Hodgkiss W S. Trackingrefractivity from clutter using Kalman and particle filters[J].IEEE Transactions on Antennas and Propagation.2008,56(4):1058-1070[3]Kotecha J H and Djuric P M. Gaussian particle filtering[J].IEEE Transactions on Signal Processing.2003, 51(10):2592-2601[4]Bolic M, Athalye A, Djuric P M, and Hong S. Algorithmicmodification of particle filters for hardware implementation.Proc. of the European Signal Processing. Conference, Vienna,Austria, 2004: 1641-1646.[5]Lin G H, Xu H F, and Masao F. Monte Carlo andquasi-Monte Carlo sampling methods for a class of stochasticmathematical programs with equilibrium constraints[J].Mathematical Methods of Operations Research.2008, 67(3):423-441[6]Wu Y X, Hu X P, and Hu D W. Comments on GaussianParticle Filtering[J].IEEE Transactions on Signal Processing.2005, 53(8):3350-3351[7]Wolfgang J. Quasi-Monte Carlo sampling to improve theefficiency of Monte Carlo EM[J].Computational Statistics Data Analysis.2005, 48(4):685-701[8]Bolic M, Djuric P M, and Hong S. Resampling algorithmsand architectures for distributed particle filters[J].IEEETransactions on Signal Processing.2005, 53(7):2442-2450[9]Wolfgang C S and Andreas U. Techniques for parallelquasi-Monte Carlo integration with digital sequences andassociated problems[J].Mathematics and Computers inSimulation.2001, 55(1-3):249-257 -
计量
- 文章访问数: 4160
- HTML全文浏览量: 73
- PDF下载量: 869
- 被引次数: 0
下载:
