Positioning System Using Classification of Ultra Short Electromagnetic Pulse Forms by ANN

The positioning system based on analysis of ultra short electromagnetic impulse forms by artificial neural network is proposed. The change of short impulse form that takes place in ultrawideband antennas of short electromagnetic pulses is used for direction determination. There is no the time synchronization between receiver and transmitter, any information transferring between them in the system. The classification of forms of the received electromagnetic impulses of different antennas is carried out by deep neural network. The positioning system is simulated for the case of plane rectangular domain irradiated by two bow-tie antennas. These antennas are excited by short Gaussian pulses of different durations. The time form of radiated fields is calculated by means of FDTD method. The analysis of the precision of the receiver coordinate determination is presented.