CAJ | 학술논문

在目标跟踪系统中会出现传感器量测无序到达融合中心的现象。考虑相同时刻过程噪声与量测噪声相关情形下的一步延迟无序量测问题，提出了一种新的滤波算法。该算法采用改进的信息滤波器，可有效处理相关噪声；采用前向预测方法，在有序状态转移方式下进行状态递推，可处理延迟量测。新算法性能不依赖于过程噪声的离散化模型，且无需对状态转移矩阵求逆。仿真实验表明，在处理无序量测时，新算法与现有的基于后向预测的处理算法表现相当，但存储量及计算量更小，实时性更好，算法是可取的。
재목표근종계통중회출현전감기량측무서도체융합중심적현상。고필상동시각과정조성여량측조성상관정형하적일보연지무서량측문제，제출료일충신적려파산법。해산법채용개진적신식려파기，가유효처리상관조성；채용전향예측방법，재유서상태전이방식하진행상태체추，가처리연지량측。신산법성능불의뢰우과정조성적리산화모형，차무수대상태전이구진구역。방진실험표명，재처리무서량측시，신산법여현유적기우후향예측적처리산법표현상당，단존저량급계산량경소，실시성경호，산법시가취적。
In target tracking system , the measurements of sensors may arrive out of sequence .A new algorithm was proposed,which can solve the one-step-lag Out-of-Sequence Measurement ( OOSM) problem when the process noise and the measurement noise are correlated at the same time .In order to deal with correlated noise,a modified information filter was introduced .A forward prediction approach was used to tackle the OOSM problem,which realized recursive estimation in ordered state transfer .The new algorithm does not require the inverse of the state transition matrix,and its performance is independent of the discretization form of process noise .The simulation shows that:compared with the existing backward-prediction based algorithms,the new algorithm has the equal performance,requires less data storage and computational cost,and has better real-time performance when used for dealing with OOSM .