Dispersion-Managed Wavelength Division Multiplexed Soliton Systems with Filtering

The use of wavelength-division multiplexing (WDM) in soliton communication systems was first proposed by Mollenauer et al. in [1]. As in conventional systems, WDM allows a substantial increase in system capacity compared to single channel systems. However, soliton interchannel collisions can induce residual frequency shifts that degrade the system performance through timing jitter. In practice, these residual shifts must be reduced, limiting the maximum frequency separation between the outermost channels. Ultimately, this restricts the possible number of channels and thus limits the transmission rate of WDM solitons systems. The use of dispersion management may substantially reduce the residual frequency shifts (and the resultant timing jitter), if the dispersion maps are chosen judiciously. In [2], we have developed a Lagrangian formulation that enables optimal dispersion maps to be chosen. Another possible technique is the use of fixed or sliding frequency filters that damp the residual frequency shifts [3], thus overcoming the restriction in the maximum number of channels. However, there are always residual timing shifts that, when added over the many collisions suffered by solitons in long-haul links, can be an important source of jitter. To further reduce this timing jitter and to raise the overall bit rate of a WDM soliton system, a combination of dispersion management and of frequency filtering may have to be used