一种多传感器信息融合中的分布模型

介绍了分布式数据融合和多传感器信息处理节点的网络分布问题,阐述了一个分布式的Kalman滤波器.这个滤波器不需中央处理单元,每个节点对全局的最佳估计是通过状态空间中心获得而不是集中拓扑.它能直接对不同结构的节点和状态空间进行通信和同化问题的相关估计,解决了内部节点的转换.这种分布式结构能从个别节点得到融合信息,任何节点都可作出对全局的估计.

The Distribution Model of a Multi-sensor Data Fusion

After a discussion over the problem of data fusion in a distributed network of multi-sensor processing nodes,the authors have formulated a distributed Kalman filter which has no central processor.Global optimum estimates at each node can be obtained by requiring an underlying centrality of the state space rather than a centralized topology.The problem of communicating and assimilating relevant estimates directly between nodes with different models and state subspaces is solved using inter-nodal transformations.Model distribution can receive the information from the specific node;any node can make the global optimum estimate to the overall situation.