Estimation of the environmental impact on the accuracy of signal recognition

Abstract

The problem of the random signals recognition system’s adaptation to the variable environmental conditions is discussed. The constructive method that demonstrates possibilities to create recognition systems able to adapt to changing working conditions. The efficiency of the method is demonstrated by experiment analyzing the recognition of random signals in environments with different characteristics. The results demonstrate that the suggested method also can be useful in the development of the speech recognition devices operating in various environments.