One Recurrent Neural Networks Solution for Passive Localization

摘要

A revised recurrent neural networks method for the passive source localization is proposed in this paper. The cost function has been designed which makes the transformation from transmitters’ localization into the outputs of the settled revised recurrent neural networks through spatial partition. The neural networks are chaotic and stable in convergence. The received signal model is constructed firstly. The parameters of the recurrent neural networks have been trained properly according to the scene. The experiments and analysis display that the revised recurrent neural networks solution not only obtains high precision location, but also has high convergence rate.