混沌映射采样的粒子滤波器

• 作 者: 宁小磊 陈战旗 赵新 何星 石国祥
• 机 构: 中国华阴兵器试验中心,第二炮兵工程学院304教研室,
• 出 处: 控制与决策
• 发表日期: 2011年12期
• 关键词: 粒了滤波 粒子滤波器 混沌映射 粒了枯竭 选粒 非线性估计 非线性系统 仿真结果 粒子退化 重采样

针对粒子滤波粒子多样性减弱引起的粒子枯竭问题,提出一种新的基于混沌映射采样的粒子滤波改进算法(CMS-PF).在重要性采样之后,用类似载波的方法将离线生成的混沌序列映射到以较大权重粒子为中心的样本子空间,从而生成一些映射粒子,并结合当前时刻的预测粒子构建候选粒子集,最终依据各粒子自身的权重实现优选.仿真结果表明,该算法能有效提高对非线性系统状态的估计精度.

Aiming at the problem of particle exhaustion caused by the decrease of particle filter particle diversity, a new particle filter improvement algorithm based on chaos map sampling (CMS-PF) is proposed.After the importance sampling, a similar carrier method is used to generate the offline chaotic sequence And then maps to the sample subspace centered on the larger weighted particles to generate some mapped particles and builds the candidate particle sets based on the predicted particles at the current time to optimize the algorithm according to the weights of the particles.A simulation result shows that the proposed algorithm is effective Improve the estimation accuracy of the nonlinear system state.