Performance enhanced DDO-OFDM system with adaptively partitioned precoding and single sideband modulation.

As a promising solution for short-to-medium transmission systems, direct detection optical orthogonal frequency division multiplexing (DDO-OFDM) or discrete multi-tone (DMT) has been intensively investigated in last decade. Benefitting from the advantages of peak-to-average power (PAPR) reduction and signal-to-noise ratio (SNR) equalization, precoding techniques are widely applied to enhance the performance of DDO-OFDM systems. However, the conventional method of partitioning precoding sets limits the ability of precoding schemes to optimize the SNR variation and the allocation of modulation formats. Thus, the precoding transmission systems are hard to reach the capacity that traditional bit-power loading (BPL) techniques, like the Levin-Campello (LC) algorithm, can achieve. In this paper, we investigate the principle of SNR variation for precoded DDO-OFDM systems and theoretically demonstrate that the SNR equalization effect of precoding techniques is actually determined by the noise equalization process. Based on this fact, we propose an adaptively partitioned precoding (APP) algorithm to unlock the ability to control the SNR of each subcarrier. As demonstrated by the simulation and experimental results, the proposed APP algorithm achieves the transmission capacity as high as the LC algorithm and has nearly 1 dB PAPR reduction. Besides, the look-up table (LUT) operation ensures low complexity of the proposed APP algorithm compared with LC algorithm. To avoid severe chromatic dispersion (CD) induced spectral fading, single sideband (SSB) modulation is also implemented. We find that SSB modulation can reach the capacity of double sideband (DSB) modulation in optical back-to-back (OB2B) configuration by optimizing the modulation index. Therefore, the APP based SSB-DDO-OFDM scheme can sufficiently enhance the performance of cost-sensitive short-to-medium reach optical fiber communication systems.