基于优化组合重采样的粒子滤波算法

重采样过程的引入,消除了粒子滤波(PF)过程中的粒子匮乏现象,使PF方法迅速地在多个领域内得到应用,但重采样过程却削弱了粒子的多样性,从而导致滤波性能下降,甚至滤波发散.提出了一种基于优化组合的重采样方法,通过选取粒子和被抛弃粒子的适当线性组合而产生新的粒子,增加了粒子多样性,从而提高PF算法的精度.仿真结果表明,步长系数为零时,该算法等价于基本的PF算法;步长系数很大时,该算法不能收敛;在适当选择步长系数的情况下,该算法的滤波性能高于基本的PF算法.介绍了该重采样算法,仿真结果证明了该算法的有效性.

A Particle Filter Algorithm Based on Optimizing Combination Resampling

The introduction of resampling process to the particle filter eliminates the degeneracy phenomenon during filtering and makes it rapidly applied in many fields,but the process leads to the loss of diversity in particles.The loss makes filter's performance worse,even makes filtering diffuse.An optimizing combination resampling(OCR) method was developed based on adaptive combination step-length,which generates new particles by combining the selected particles and abandoned particles.Particle filter(PF) with the new method partly overcomes the loss of diversity in particles,and improves its precision.When the step-length is set zero,the algorithm is equal to PF;when the step-length is great,the algorithm diffuses;and when the step-length is properly set,the algorithm can achieve better performance than PF.The paper put its emphasis on the new resampling algorithm,and then did some simulation experiments to give proof.The simulation results prove the effectiveness of the proposed method.