Ultra-wideband beamforming by using a complex-valued spatio-temporal neural network
Suksmono A.B.a, Hirose A.b
a Department of Electrical Engineering, Institut Teknologi Bandung, Indonesia
b Department of Electrical and Electronics Engineering, University of Tokyo, Japan
Abstract
We propose an ultra-wideband (UWB) beamforming technique by using a spatio-temporal complex-valued multilayer neural network (ST-CVMLNN). The complex-valued backpropagation through time (CV-BPTT) is employed as a learning algorithm. The system is evaluated with an ultra-wideband monocycle signal. Preliminary simulation results in suppression of UWB interferer and in steering for desired UWB signal, demonstrating the applicability of the proposed system. © Springer-Verlag Berlin Heidelberg 2004.
Author keywords
Back-propagation through time,Beamforming technique,Complex-valued,Spatio temporal,UWB signals,Wideband beam forming