摘要
在非线性滤波系统中,线性化误差和系统误差是影响滤波精度的两个主要因素.提出联合粒子滤波方法,在减少线性化误差的同时,能够实时估计系统误差的大小并自适应地消除其影响.说话人跟踪仿真实验结果验证了此方法的有效性.
In nonlinear filtering system, linearization error and system error are two major factors that affect filtering accuracy. A joint particle filtering method is proposed to estimate system error in real time and eliminate its effect adaptively as well as to minimize the linearization error. Simulation results of speaker tracking demonstrate the effectiveness of the proposed method.
出处
《大连交通大学学报》
CAS
2007年第4期59-63,共5页
Journal of Dalian Jiaotong University
关键词
系统误差
联合滤波
粒子滤波
线性化误差
system error
joint filter
particle filter
linearization error