Adaptive channel identification in optical fiber communication systems

In this paper, adaptive methods for the identification of the channel response of optical fiber communications links are considered, in the context of intensity modulated direct detection on/off keying transmission. A discrete time, second order Volterra kernel is used for the modeling of the overall channel response. The Volterra parameters are estimated on the basis of the available input/output data, by resorting to adaptive system identification techniques. The performance of the adaptive processing algorithms is evaluated in terms of the achieved minimum estimation error, the rate of the convergence and the required computational complexity. It is demonstrated, by means of computer simulation, that the use of low cost adaptive processing algorithms, properly tailored to the problem at hand, can provide a reliable solution, without sacrificing much of the performance.