Estimating Markov Model Parameters from Statistical Analysis of Speech Packets Transmitted over BWA Networks

Abstract

The main problem encountered in the design of Markov models is in choosing the adequate mathematical model for their representation. Analysis of the mathematical models of speech packet sources was conducted on the basis of Markov chains. The specific characteristics of telephony dialogue and monologue were taken into consideration. These in their turn can be used in the design of methods of statistical companding of packet communication nodes. In this paper we show that the choice of the number of states of the Markov chain as well as their probabilistic characteristics can be estimated from the results of the statistical processing of speech packets transmitted over Broadband Wireless Access Networks (BWAN).

Keywords: Markov models, Wireless network, Speech processing, Packet switching, Mathematical model.