A Rough Neural Network Algorithm for Multisensor Information Fusion

The multisensor information fusion is a key issue for multisensor system. One of its difficulties lies in the switching of the state of sensor clusters. That is, which direction should the sensor information been fused into at a given moment? An algorithm of multisensor information fusion based on rough set and neural network was proposed in this paper. Firstly, the typical clustering distributions of 54 sensors within one day were regarded as sample space. The rough set was used for access of knowledge to make the decision table of the "data-fusion distribution". Next, the redundant properties and samples of information in one month were removed using the method of knowledge reduction of rough set. Then, the neural network was applied for clustering and analyzing to form the distribution rules of multisensor information fusion. Finally, the rough neural fusion algorithm, the neural quotient space fusion algorithm and word computing fusion algorithm are simulated and analyzed. The results show that the model and algorithm proposed in the paper are efficient in classification and rapid in sensor clustering distribution decide.