An Adaptive Clock Recovery Algorithm for Differentiated Roll-Off Characteristics in Gridless Optical Network

In elastic optical networks, multi-granularity channels with different bandwidth impose higher demands on the digital signal processing (DSP) algorithms of receivers. For regular DSP approaches, the clock recovery algorithm is sensitive to data rate. Traditional Gardner algorithm and its modification methods struggle to dynamically adapt the different roll-off factors based on the data rate to approach the Shannon limit. To enhance the clock recovery performance for optical signals with different rates to improve spectral efficiency, this paper proposes a clock synchronization method based on an adaptive module to fit different roll-off factors. By designing a multi-channel roll-off factor loop algorithm, the method dynamically adapts to different signals transmitted over various channels. Then, we proposed an adaptive clock recovery algorithm and conduct simulations via MATLAB to show the proposed approach can decrease the timing error jitter and improve convergence speed. In high roll-off systems, the proposed algorithm increases the convergence speed by 12%, and, in low roll-off systems, decreases the timing error jitter by about 52% and 33%.