A novel algorithm for wideband DOA estimates based on neural network

Direction of arrival estimation of mobile stochastic electromagnetic sources with variable radiation powers using hierarchical neural model

“…1,[10][11][12] Artificial neural networks (ANNs) [13][14][15][16] represent an another direction of research for solving the DoA problems in real-time. [16][17][18][19][20][21][22][23][24][25][26][27][28][29][30][31][32][33][34] Beside ANNs, the more general methods in the field of artificial intelligence, such as genetic programming (GP) and support-vector machines (SVM), are used in today research for the same purpose, as demonstrated for the DoA estimation problem solving in Reference 35 and Reference 36, respectively. Although the GP and SVM-based DoA models have more generalization capabilities than the ANN models in modeling highlynonlinear dependences present in DoA problems, DoA ANN models still represent a good alternative.…”

“…Multilayer perceptron (MLP) neural networks were also used for the DoA estimation of deterministic and stochastic EM sources. [24][25][26][27][28][29][30][31][32] Regarding the stochastic scenario, in References 26-29 uncorrelated mobile stochastic sources were considered and it was shown that only the values of elements of the first row of spatial correlation matrix at the receiver were sufficient to develop an accurate neural model. In References 30 and 31, neural models were developed for the DoA estimation of partially correlated stochastic sources using not only the elements of the first row of the spatial correlation matrix, but also the elements from the upper triangle of this matrix.…”