Page 516 - Fiber Optic Communications Fund
P. 516

11






           Digital Signal Processing






           11.1  Introduction

           The key component that revived coherent fiber communications in the mid-2000 was high-speed digital signal
           processing. In the 1990s, coherent receivers used optical phase-locked loops (OPLL) to align the phases and
           dynamic polarization controllers to match the polarization of the received signal with that of the LO. However,
           dynamic polarization controllers are bulky and expensive [1], and each channel of a WDM system needs a
           separate polarization controller. Phase locking in the optical domain using OPLL is difficult as well. With
           the advances in high-speed DSP, phase alignment and polarization management can be done in the electrical
           domain, as discussed in Sections 11.5 and 11.7, respectively. Linear impairments such as chromatic dispersion
           (CD) and polarization mode dispersion can be compensated using equalizers, as discussed in Sections 11.6
           and 11.7, respectively. It is also possible to compensate for the interplay between dispersion and nonlinearity
           by using digital back propagation (DBP), in which the nonlinear Schrödinger equation is solved for a virtual
           fiber whose signs of dispersion, loss, and nonlinear coefficients are opposite to those of the transmission fiber.
           DBP is discussed in Section 11.8.



           11.2  Coherent Receiver
           Fig. 11.1 shows a schematic of the coherent IQ receiver with digital signal processing. The in-phase and
           quadrature components of the received signal can be written as (see Chapter 5, Eqs. (5.114) and (5.118))
                                    y = RA A  Re{s(t) exp [−i( t +Δ)]}∕2,               (11.1)
                                                            IF
                                          r LO
                                     I
                                   y = RA A   Im{s(t) exp [−i( t +Δ)]}∕2,               (11.2)
                                          r LO
                                    Q
                                                            IF
           where s(t) is the transmitted data:
                                                 ∑
                                            s(t)=   a g(t − mT ),                           (11.3)
                                                     m       s
                                                  m
           T is the symbol period, and g(t) represents the pulse shape. Eqs. (11.1) and (11.2) have to be modified to take
            s
                                             ∘
           into account the noise and delays due to 90 hybrids:
                               y = KRe{s(t −  ) exp [−i(2f (t −  )+Δ)] + n(t)},     (11.4)
                                I
                                                              I
                                                        IF
                                             I
                               y = KIm{s(t −  ) exp [−i(2f (t −  )+Δ)] + n(t)},     (11.5)
                               Q
                                                              Q
                                             Q
                                                        IF
           Fiber Optic Communications: Fundamentals and Applications, First Edition. Shiva Kumar and M. Jamal Deen.
           © 2014 John Wiley & Sons, Ltd. Published 2014 by John Wiley & Sons, Ltd.
   511   512   513   514   515   516   517   518   519   520   521