基于DSP的改进粒子滤波算法硬件实现

粒子滤波算法在处理非线性、非高斯系统方面有明显的优势，但较大的运算量限制了其实时性的应用。针对这一问题，文中研究利用DSP处理器高速运算的特点实现粒子滤波算法。文中针对粒子滤波的粒子退化问题给出重采样方法，详细给出了DSP实现粒子滤波算法的流程，对DSP的存储资源进行了优化设计。利用典型模型进行算法实现验证，结果表明：采用DSP实现的粒子滤波算法可对状态参量进行精确处理，验证了DSP用于非线性、非高斯系统的粒子滤波算法实现的有效性。

Implementation on Improved Particle Filter Based on DSP

Particle filter algorithm has great advantages in dealing with nonlinear and non-Gaussian system. However,the huge amount of calculation limits its application in real time.To solve this problem, this paper propose a method by using the high-speed operation characteristics of the DSP to implement the particle filter algorithm. This paper use the resampling method to solve the particle degradation and give out the detailed flow of the particle filter algorithm implemented by DSP and optimize storage resources of the DSP. Using the typical model to demonstrate the performance of DSP implementation, the simulation results show that the designed particle filtering algorithm implemented by DSP can process exactly the state parameter and the implementation of particle filter by DSP is effective.