摘要
标准的卡尔曼滤波算法由于使用了固定单一的状态噪声模型,因此当目标运动状态经常发生剧烈变化时,跟踪效果不是很理想。为了提高对目标的跟踪精度和跟踪收敛速度,提出了一种新的算法,通过多次步长不同的卡尔曼滤波算法来判断机动目标的运动状态,进而使得系统状态噪声协方差能够随着目标机动情况自适应调整。最后的蒙特卡罗仿真实验验证了此算法的有效性。
Because the state noise model is fixed and simplex in kalman filter, the result of tracking is unsatisfactory while the moving state of target changing acutely. In order to improve the tracking precision and convergence rate, a novel algorithm was proposed, multi-kalman filter with different sample time were used to judge the target moving state, and then adjust the state noise covariance adaptively. From the simulation, it can be seen that the algorithm proposed can improve the result of tracking.
出处
《系统仿真学报》
EI
CAS
CSCD
北大核心
2008年第23期6458-6460,6465,共4页
Journal of System Simulation
关键词
目标跟踪
卡尔曼滤波
收敛性
自适应
仿真分析
target track
kalman filter
astringency
adaptive
simulation analyses
