多假设最大联合概率决策融合准则的自适应算法

Adaptive decision fusion for multiple hypotheses maximum joint probability decision rule

为获得多假设最大联合概率决策融合准则的多假设自适应算法,通过对原算法中各权重系数求偏微分,结合简单计数方法,给出了自适应决策融合中的各个系数增量表达式,并进行了收敛性证明.该算法在系统参数未知及参数发生改变的情况下,均能进行决策融合.针对两假设自适应决策融合问题,将该算法与Mansouri的算法进行了比较.最后在多假设情况下对该算法进行了应用,并给出了提高收敛速度的方法.

To obtain the corresponding adaptive decision fusion methodology for multiple hypotheses based on the maximum joint probability （MJP） fusion rule, by using partial differentiation involving with simple counting technology, the increment expression of each coefficient in the rule is given, and its convergence is proved. The algorithm derived can deal with the decision fusion system under the condition that the parameter is unknown or changed varied with time. Comparing with classical algorithm, the adaptive algorithm shows its superiority, under the condition of dealing with an two hypothesis adaptive decision fusion case. In the multiple hypothesis case study, the effectiveness of proposed algorithm is indicated, and the fast convergence strategy is also developed and applied to the adaptive decision fusion.