论文部分内容阅读
针对常规自适应网格多模型滤波方法存在的模式跟踪能力不强,在系统模式发生大幅度跳变时,滤波模型集难以快速适应模式变化,容易出现模式跟踪失败的问题,模拟生物免疫系统识别抗原的原理,提出一种新的多模型滤波方法.该方法在原有算法基础上,增加了一个监测模型集来随机探测当前模式的分布信息,并对滤波模型集的适应过程进行引导和纠偏.仿真结果表明该方法能够获得更好的模式跟踪能力.
For the traditional adaptive multi-model filtering method, the existing model tracking ability is not strong. When the system mode changes dramatically, the filtering model set can not adapt to the mode change easily, and the model tracking failure is easy to occur. Simulating biological immune system recognition Antigen principle, a new multi-model filtering method is proposed.This method adds a monitoring model set based on the original algorithm to randomly detect the distribution information of the current mode and guides and rectifies the adaptation process of the filtering model set. The simulation results show that this method can get better mode tracking ability.