基于量化新息的容积粒子滤波融合目标跟踪算法

Target Tracking Algorithm Based on Cubature Particle Filtering Fusion with Quantized Innovation

针对现有非线性网络化目标跟踪融合算法存在的精度低和实用性差等不足,以一类带有噪声相关的非线性网络化目标跟踪系统为对象,研究基于测量新息量化策略和容积粒子滤波(Cubature particle filter,CPF)的目标跟踪融合算法.首先,利用状态方程恒等变换和矩阵相似变换理论解除过程噪声与测量噪声以及测量噪声之间的相关性;其次,各个传感器节点采用自适应策略量化局部测量新息并将其发送到融合中心(Fusion center,FC);随后,在集中式融合框架下采用容积粒子滤波器设计基于测量值扩维的量化融合跟踪算法,进而给出相应的顺序滤波量化融合算法,上述算法可有效解决因自适应量化引起的非高斯问题;最后,通过两个计算机仿真实验验证了所提出跟踪算法的有效性.

For the nonlinear networked target tracking system with arbitrarily correlated noises, target tracking fusion algorithms based on quantized innovation and cubature particle filter （CPF） are researched in order to overcome the shortages of the existing methods, which have low precision and poor real-time performance. Firstly, identical transformation of state equation and matrix similarity transformation theory are used for arbitrary noises decorrelation. And then, each sensor node adopts an adaptive quantization strategy to quantize its innovations, and transmits them to the fusion center （FC）. Subsequently, the nom-Gaussian problems caused by quantization are solved by using CPF to design a fusion tracking algorithm with augmented measurements in the centralized fusion framework. Moreover, its corresponding sequential fusion form is developed. Finally, two computer simulation experiments show the effectiveness of the proposed method.