CUSUM application in perceptual speech quality control

One of the important aspects of the mobile communication industry is satisfying customers' needs most economically. Indeed, customers expect good and consistent quality of service from the provider. In mobile telephony, this amounts to controlling the speech quality as “perceived” by the customers. Controlling the perceptual speech quality necessitates reliable measurement of the quality first, followed by exercising direct control of it. The end-users' quality has traditionally been monitored and controlled based on radio link measurements such as the Signal Interference Ratio (SIR), Bit Error Rate (BER), or Frame Error Rate (FER). The ultimate measure of perceived speech quality is realized through subjective listening tests, but this is not practical for real-time day to day applications. In recent years, objective quality measurement algorithms have been developed to predict the subjective quality with considerable accuracy. The state of art in the International Telecommunication Union's Telecommunication Standardization sector (ITU-T) recommendation for referenced objective quality measurement method is P.862 Perceptual Evaluation of Speech Quality (PESQ) model. These algorithms have yet to be applied for end user quality control in cellular networks using Statistical Process Control (SPC). In this paper, the application of Cumulative Sum (CUSUM) scheme, based on the PESQ algorithm for perceptual speech quality control is presented.