摘要
将序列概率比检验(SPRT)和期望极大化(EM)算法相结合,给出了一种解决杂波环境下目标航迹起始问题的新方法。首先,利用两点差分方法建立暂时航迹;然后,针对由于目标真实状态未知导致的应用SPRT算法时似然函数计算上的困难,导出了基于递推EM算法的似然函数计算公式,可以实时且有效地解决其中的不完全数据问题。从而,实现暂时航迹的确认和删除。仿真结果表明文中算法具有良好的性能。
A new method for track initiation in clutter environment is presented,which combines the sequential proba-bilistic ratio test(SPRT)an d expectation maximization(EM).Firstly,the two points difference procedure are adopted to form the temporary tracks.Secondly,aiming at the computational difficulty of the likelihood function due to the unknown target state,the recur sive EM-based formula of likelihood function is introduced which can provide a real-time and ef-fective solution to the incomplete data problem.As a resul t,the temporary tracks are confirmed or rejected.The simulation results indica te that the proposed method performs well.
出处
《计算机工程与应用》
CSCD
北大核心
2003年第14期66-69,共4页
Computer Engineering and Applications
基金
国家重点基础研究发展规划项目(973)(编号:2001CB309403)
关键词
航迹起始
序列概率比检验
期望极大化
track initiation,sequent ial probabilistic ratio test(SPRT),expectation maximization(EM)
