NEURAL MODEL FOR FAR-FIELD 1D LOCALIZATION OF MOBILE STOCHASTIC EM SOURCES WITH PARTIALLY CORRELATED RADIATION




Abstract:
This paper considers a possibility of angle position determination of mobile stochastic sources with partially correlated radiation, where antenna array and multilayer perceptron neural network processing are used. It is shown that the neural model trained with samples from a system with uncorrelated source radiation cannot determine position of sources with a satisfactory accuracy when sources have some degree of correlation in radiation. That is why it is suggested training samples to be generated for different values of partial source correlation. This kind of generated neural network training may provide source position determination with satisfactory accuracy even when there is partial correlation, which in the paper is presented with an example of two sources that linearly move in azimuth plane

CITATION:

IEEE format

Z. Stanković, N. Dončov, I. Milovanović, M. Sarevska, B. Milovanović, “Neural Model for Far-Field 1D Localization of Mobile Stochastic EM Sources with Partially Correlated Radiation,” in Sinteza 2017 - International Scientific Conference on Information Technology and Data Related Research, Belgrade, Singidunum University, Serbia, 2017, pp. 169-175. doi: 10.15308/Sinteza-2017-169-175 

APA format

Stanković, Z., Dončov, N., Milovanović, I., Sarevska, M., Milovanović, B. (2017). Neural Model for Far-Field 1D Localization of Mobile Stochastic EM Sources with Partially Correlated Radiation. Paper presented at Sinteza 2017 - International Scientific Conference on Information Technology and Data Related Research. doi:10.15308/Sinteza-2017-169-175

BibTeX format
Download

RefWorks Tagged format
Download