Computer Science · Artificial Intelligence
The Distribution Model of a Multi-sensor Data Fusion
See on the mapAbstract
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.