粗糙集理论的上下近似集合可以刻画传感器对未知雷达辐射源信号的识别程度。从灰关联理论对传感器测得的原始数据计算出的不肯定度出发,计算每个目标与传感器测得的悲观解和乐观解距离。从而计算出目标与悲观解和乐观解间的灰关联系数矩阵。用假定的折中系数α、β计算出各传感器的侦察权重。再由灰关联系数矩阵和传感器权重得到各雷达辐射源信号的与悲观解和乐观解在传感器Sj识别意义下的灰关联度,最后计算融合判别函数得到融合结果。仿真实验表明,由粗糙集理论和灰关联理论相结合的多传感器数据融合的方法可以有效地应用于雷达辐射源信号识别,特别在是在降低多传感器引起的不确定性取得较好的效果。
Recognition degree of unknown radar emitter signal can bedesefibe by upper and lower approximation sets of rough sets. From the uncertain degree measure by multi-sensors, pessimism and optimism distances of every object to sensors and grey association matrix can be calculated by grey association theory. According α/β and grey association matrix, grey association degree and fusion result can gain. The simulation results show the method of combining rough sets and grey association theory is effective, and it can be applied on radar emitter signal recognition, especially in decreasing the uncertain degree introduced by multi-sensors