基于Kohonen网络的动目标实时定位融合算法

Moving Target Real-Time Location Fusion Algorithm with Kohonen Network

为完成对动目标的实时定位测量,有效减小测控装备组网测量实时跟踪定位误差,在建立多定位子系统组网混合式融合结构模型的基础上,利用Kohonen网络的自组织特征映射模型特点,确定各组交会数据的权值,并将其输入1个Kohonen网络进行在线竞争、融合求得目标的位置信息,增强了目标实时位置信息的连续性、完整性,减小了实时定位误差。

To facilitate real-time location measurement of moving targets and reduce target location error during real- time tracking with a networked measurement system, the weight factor of each group is determined by setting up a real-time hybrid structural model network with multiple location subsystems and by making use of the characteristics of the self-organizing nerve of the network. The weight ~actor is then inputted into a Kohonen network to compete with each other online to get the location information of the target. As a result, the continuity and integrity of the target real-time location information is enhanced and real-time positioning errors are reduced.