An Improved Kramers-Kronig Receiver with Upsampling-free and Low Carrier-to-Signal Power Ratio

The Kramers-Kronig (KK) receiver breaks through the limitation that a single photodetector (PD) can only process the intensity information, and has become a research hotspot in optical communication system. However, in the conventional KK algorithm, the required nonlinear operations (such as logarithm (log) and exponential and square-root (sqrt) functions) significantly broadens the signal spectrum, it is necessary to employ the digital upsampling at the beginning of the digital signal processing (DSP). The high sampling rate will bring multiple hardware resources and power consumption, which becomes a major obstacle to implement KK receiver. In this paper, we propose a novel improved algorithm for the KK receiver, which does not need the digital upsampling and lowers the required carrier-to-signal power ratio (CSPR) with a comparison of the typical upsampling-free KK receiver. In the proposed improved algorithm, we adopt the mathematical approximations of Taylor's expansion and Newton's tangent method to avoid the use of log and sqrt functions. Then we removed the exponential function by expressing complex signal in Cartesian form (i.e., real plus imaginary). Whereas the real part of it is induced by employing imaginary part-induced signal-signal beat interference (SSBI) removal and sqrt operation to recover. Meanwhile, we use the simplified hybrid KK-SSBI cancellation (SHKK-SSBIC) technology to reduce the required CSPR by the system. We validate our proposed algorithm by transmitting 160 Gb/s single-sideband (SSB) signal. The experimental results show that compared with the upsampling-free KK receiver, the proposed scheme can realize the reduction of the required CSPR by 1 dB and 0.8 dB respectively under the Nyquist sampling rate at (back-to-back) BTB and 80 km transmission. Moreover, our proposed method achieves a 2 dB system sensitivity improvement in the BTB scenario. We also discuss the hardware implementation of the improved KK algorithm and its computational complexity.