Recognition and Identification by Timbre of a Living Creature's Voice Based on Trainable Neural Networks in Real-Time Mode and Their Implementation in the Intelligent System «Neurocyber»

Abstract

The article considers physical basis for formation and propagation of sound and its psychophysical perception by living creatures. The authors suggest an algorithm of living creatures’ identification by the voice timbre and the algorithm’s practical realization by artificial neural networks, by initialization of a conversation by the robot in order to reduce ambiguity of responses and to improve the quality of recognition in real-time mode.