Abstract
Conventional fast Fourier transform based frequency offset estimation (FFT-FOE) algorithm is suitable for QPSK and 8/16/64QAM signals. However, due to the non-rectangular distribution of constellation points, the conventional FFT-FOE is not suitable for 32-QAM signal. In this paper, we report a factor $\gamma $ which can indicate whether the FFT-FOE algorithm can be used for frequency offset estimation for different modulation formats. The FFT-FOE algorithm is hard to apply for frequency offset estimation if the value of $\gamma $ is low. Fortunately, the value of $\gamma $ can be increased by digital amplification and phase rotation. Based on the digital amplification and phase rotation, a modified FFT-FOE algorithm is proposed and can realize the frequency offset estimation of 32-QAM signal with only 512 symbols. The bit error ratio (BER) of a 28 Gbaud 32-QAM signal is lower than the soft-decision forward-error correction (SD-FEC) limit with the optical signal-to-noise ratio (OSNR) of 22 dB. The proposed method has better robustness to phase noise. If BER at SD-FEC limit is considered for 28Gbaud 32-QAM signal, the required OSNR is relaxed by 2 dB when the laser linewidth is greater than 5 MHz. This will make sense when the wide-linewidth laser is used in short reach optical communication system. The experimental results from 10 Gbaud 32-QAM system with 300 km fiber transmission indicate that the proposed method can be applied when OSNR is greater than 14 dB. The results demonstrate that the proposed frequency offset estimation scheme is also suitable for QPSK and 8/16/64QAM signals. Thus, the proposed algorithm can perform blind frequency offset estimation for a coherent transceiver.
Highlights
With the development of the internet and the increasing demand for IP traffic, optical networks have gradually evolved from fixed architectures to flexible architectures, which can increase spectral efficiency by modulation format switching [1]–[3]
The experimental results from 10 Gbaud 32-quadrature amplitude modulation (QAM) system with 300 km fiber transmission indicate that the proposed method can be applied when optical signal-to-noise ratio (OSNR) is greater than 14 dB
We propose a digital amplification and phase rotation based fast Fourier transform based frequency offset estimation (FFT-FOE) algorithm for coherent transceiver, which is transparent to quadrature phase shift keying (QPSK) and 8/16/32/64QAM signals
Summary
With the development of the internet and the increasing demand for IP traffic, optical networks have gradually evolved from fixed architectures to flexible architectures, which can increase spectral efficiency by modulation format switching [1]–[3]. A new frequency offset estimation algorithm that can be applied before modulation format identification for coherent transceiver is meaningful. Since the fourth-order constellations of QPSK and 8/16/64QAM signals are non-circular, conventional fast Fourier transform based frequency offset estimation (FFT-FOE) can be used for identifying respective frequency offsets. Due to the non-rectangular distribution of constellation points, the conventional FFT-FOE cannot be applied to the 32-QAM signal with 512 symbols unless there are huge number of symbols. To solve this problem, a FFT-FOE technique by selecting and digitally amplifying the inner QPSK ring of 32-QAM signal was reported in [16]. We propose a digital amplification and phase rotation based FFT-FOE algorithm for coherent transceiver, which is transparent to QPSK and 8/16/32/64QAM signals. The experimental results from 10 Gbaud 32-QAM system with 300 km fiber transmission indicate that the proposed method can be applied when OSNR is greater than 14 dB
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
