多传感器多目标滤波算法中的计算问题探讨

Discussion on calculation of multi-sensor multi-target filtering algorithm

多传感器多目标跟踪已广泛应用于军事、工业、农业及人们的日常生活之中。多目标跟踪是将传感器接收到的众多量测值按照来源分类,给出各跟踪目标的运动轨迹,并分析运动轨迹的准确性和可靠性。该文利用有限统计理论(FISST)、高斯型单目标、多目标滤波近似似然函数,对多目标滤波算法中存在的精度计算问题进行研究,为多传感器多目标的检测、数据融合及预测提供了一个简化、全概率运算依据。

Multi -sensor multi -target tracking has been widely used in military, industry, agricul- ture and people＇s daily life. Multiple target tracking is many received sensor measurements in accordance with the source classification, the trajectory tracking, and motion trajectory analysis accuracy and reliabil- ity. This paper uses finite statistical theory （FISST）, Gauss type single objective, multi -objective filter approximate likelihood function, to study the calculation problems of multi objective filtering algorithm, for multi - sensor multi - target detection, data fusion and prediction provides a simplified, the probabili- ty calculation basis.