A Neural Network Based Approach For Sensors Issued Data Fusion
In this paper, we present a Functional Link Network (FLN) based neural technique for sensors issued data fusion. Thanks to a pruning algorithm, we build dynamically the internal layer of the FLN, constructing an optimal architecture of the FLN neural network defining an optimal fusion policy. As the neural FLN minimize the mean square error (MSE) during the learning step, an optimal fusion policy is reached in the sense of the MSE. Experimental results related to performances enhancement of metric sensors have been reported validating our approach.
KeywordsMean Square Error Data Fusion Radial Basis Function Pruning Algorithm Optimal Architecture
Unable to display preview. Download preview PDF.
- 1.I. Bloch (1996), “Information Combination Operators for Data Fusion: A Comparative Review with Classification”. IEEE Transaction On SMC, part A,Vol. 26, N°1, pp 5267, 1996.Google Scholar
- 2.Y. Cheng and R. L. Kashyap (1992), “Recursive Fusion Operators: Desirable properties and illustrations”. Data Fusion in Robotics and Machine Intelligence“, Ed. M. A. Abidi and R. C. Gonzalez, Academic Press inc, 1992.Google Scholar
- 3.Brown M, Harris C.J, “Neural Networks for Modeling and Control” in Advances in Intelligent Control, Taylor Francis Lts, pp. 17–57, ISBN 0–7484–0066–4.Google Scholar
- 4.A. Sachenko, V. Kochan, V. Turchenko, V. Golovko, J. Savitsky, A. Dunets, T. Laopoulos (2000), “Sensor errors prediction using neural networks”, Proceedings of the IJCNN’2000 Conf., Jul 24-Jul 27 2000, 2000, Como, Italy, p 441–446.Google Scholar
- 6.Nageswara S. V. Rao (1997), “Fusion Rule Estimation Using Vector Space Methods” Proceedings of SPIE, Sensor Fusion: Architectures, Algorithms, and Applications, volume 3067, pp 130–135, 24–24 April 1997, Orlando, Florida.Google Scholar
- 7.A. Sachenko (1999), “IMCS development on the basis of distributed sensor networks”, IEEE AFRICON Conference, vl, 1999 IEEE, Piscataway, NJ, USA, p 345–350.Google Scholar