Photonic MMW Real-Time System: FPGA-Enabled Low-Complexity Frequency Offset Estimation and Symbol Synchronization

Photonics-aided millimeter-wave communication (PMC) offers the advantage of wide bandwidth, enabling the integration of fiber-wireless links for ultralong-distance, high-capacity transmission. However, due to the temperature drift and large linewidth of laser sources, photonics-aided millimeter-wave transmission often requires more complex digital signal processing (DSP) to manage frequency offsets of several hundred megahertz and residual phase offsets of several hundred kilohertz, which significantly increases the resources of the receiver-end DSP. Moreover, in real-time communication scenarios, the sampling frequency offset (SFO) caused by the mismatch between the transmitter and receiver rates affects the symbol synchronization (SS) at the receiver and the performance of the adaptive filter. In this article, we propose an optimized DSP architecture for the photonics-aided millimeter-wave backend, where a novel SS scheme is implemented to address the issue of transceiver-rate mismatch. This scheme leverages the tracking characteristics of a fractional adaptive equalizer. Furthermore, we design a novel frequency offset estimation (FOE) architecture by utilizing phase ambiguity in conventional phase offset estimation algorithm, and we present a variant of this architecture for higher order modulation formats. Both novel SS and FOE method are modulation-format compatible. We have completed the performance evaluation of the new SS and FOE methods in simulations. Finally, we have implemented full-process real-time receiver DSP flow on field-programmable gate array (FPGA), demonstrating real-time equalization and recovery of photonics-aided millimeter-wave signals. This system is experimentally validated and successfully achieved 4.6-km antenna polarization multiplexing photonics-aided millimeter-wave wireless transmission.