A recursive algorithm for maximum likelihood trend estimation in exponential noise and its application to IP telephony

A recursive algorithm for maximum likelihood estimation of clock skew i.e. difference between the transmitter and receiver clock frequencies in telephony over asynchronous packet switched networks e.g. Internet telephony, is derived. The underlying maximum likelihood method [9] is based on modeling the clock skew as linear trend and the network delay jitter as an i.i.d. random process with exponential distribution. The memory requirements of the proposed algorithm are data dependent and are investigated by simulations using both computer generated and recorded data sets.