Robust Channel Estimation for Coherent Optical Orthogonal Chirp-Division Multiplexing With Pulse Compression and Noise Rejection

Orthogonal chirp-division multiplexing (OCDM) was recently proposed as a promising modulation technique for high-speed fiber-optic communications due to its resilience to transmission impairments. By exploiting the pulse-compression property of the chirped waveforms, in this paper we propose a channel estimation algorithm compatible with the coherent optical (CO) OCDM system. In the proposed algorithm, a chirp from the orthogonal basis of the Fresnel transform is employed as the pilot signal. At the receiver, noise-rejection windowing is adopted in the Fresnel domain to remove the excessive noise after pulse compression, and then dispersions can be effectively compensated by efficient single-tap equalizers. It is shown that the proposed estimator is unbiased as long as the width of the noise-rejection window is wider than the maximum excess delay of the system. Moreover, the optimal window function achieving the minimum mean square error is derived in closed form. With these desired properties, it is shown that the proposed estimator outperforms conventional estimators, such as the intrasymbol frequency-domain average estimator, especially when the dispersion is severe. Finally, numerical results are provided to confirm the advantages of the CO-OCDM system with the proposed channel estimation algorithm.