A fuzzy logic-based speech detection algorithm for communications in noisy environments

In the field of mobile communications correct Voice Activity Detection (VAD) is a crucial point for the perceived speech quality, the reduction of co-channel interference, the power consumption in portable equipment. This paper shows that a valid alternative to deal with the problem of activity decision is to use methodologies like fuzzy logic, which are suitable for problems requiring approximate rather than exact solutions, and which can be presented through descriptive or qualitative expressions. The Fuzzy Voice Activity Detector (FVAD) proposed uses the same set of parameters adopted by the VAD in Annex B to ITU-T G.729 and a set of six fuzzy rules automatically extracted through supervised learning. Objective and listening tests confirm a significative improvement respect the traditional methods above all for low signal-to-noise ratios.