Attribute fusion based on NFE model and convex optimization approach
Mei Liu, T. Quart, Lei Zhao · 2004
Due to the drawbacks of D-S evidential inference theory approach such as difficulty of obtaining basic probability assignments, the need of being exhaustive and exclusive evidence of the reports, and a computational complexity suffering from an exponential growth, an attribute fusion structure based on NFE (neural fuzzy expert) model and convex approach is presented. This paper also presented how to make sensor reports with fuzzy expert system and acquire the reliability of sensors with fuzzy neural network. The method to fusion sensor reports with convex optimization approach is also discussed. This method makes good use of the logic reasoning ability of fuzzy expert system and the self-learning, self-reasoning ability of fuzzy neural network. The fusion results show that these methods can fusion efficiently in real time with the more reliable results.