Study of Information Fusion Evaluation Methodology Based on Multiple Approximation Space

The theories and methods of performance evaluation for multi-sensor data fusion system based on Multiple Approximation Space are discussed. A information fusion performance criteria system is founded with C3I applications as background, the evaluation process is similar to the decision process, the general way of integration system evaluation and the algorithm of assign the weights to condition attributes using rough set are presented. In the meantime, a framework is presented for performance test and evaluation of the multi-sensor data fusion. At last, the algorithm of information fusion and the choice of parameters can be directed by using software, and experimental results reveal the efficiency of the method.