摘要
在目标跟踪过程中,强跟踪滤波可以对目标的突变状态进行较好的跟踪,但由于系统噪声和观测噪声协方差阵是时变的未知量,采用固定数值将会影响目标跟踪的精度,所以需要对两种噪声特征进行在线估计。应用噪声有限记忆自适应算法改进强跟踪滤波,在线自适应估计噪声协方差阵,并提出基于时变噪声系统自适应强跟踪滤波的目标预测算法,提高了对目标状态的估计精度。利用自主研发的全自主足球机器人进行目标跟踪的仿真,结果验证了该算法的有效性。
In the process of target tracking, the strong tracking filter is sensitive to mutational states. Because the covariance matrixes of system noise and measurement noise are time-varying unknown quantities, using fixed value will affect the accuracy of tracking. So it is necessary to estimate the noise characteristics online. With the online adaptive estimation for the noise covariance matrixes, the adaptive algorithm for limited-memory noises is used to modify the strong tracking filter. Based on the adaptive strong tracking filter for the systems with time-varying noises, the target prediction algorithm is proposed so as to improve the estimation accuracy. The self-developed autonomous soccer robot is used for simulation, and the results prove the validity of the prediction algorithm.
出处
《航天控制》
CSCD
北大核心
2009年第1期3-6,共4页
Aerospace Control
关键词
自适应算法
强跟踪滤波
噪声有限记忆
全自主足球机器人
Adaptive algorithm
Strong tracking filter （ STF ）
Limited -memory noise
Autonomous soccer robot
