Time-optimised real-time time–frequency measurement technique for RoFSO applications

Hardware-based embedded solution for the time–frequency (T–F) measurement/imaging is essential nowadays for numerous applications. Several solutions exists based on the software algorithms which are more time-consuming and hence not appropriate for real-time applications. In this work, an accurate T–F measurement/imaging algorithm based on the Wigner-Ville distribution (WVD) technique is proposed and implemented in a Xilinx Virtex-7 VC709 FPGA. A time/resource/space optimised pipelined-parallel digital architecture is designed for real-time applications. The proposed digital architecture is capable of completing (with seven advantages) all the computations and measuring the T–F values within $99.38\mu s$. The designed architecture requires just 16 parallel modules, 26% of logic cores, 16% of DSP blocks, and 6% of memory. The computation performance & measurement accuracy of the same is experimentally validated with a data decoding scheme of an in-house-built radio over free space optical (RoFSO) communication system. The proposed T–F measurement algorithm, digital architecture design approach, construction of RoFSO communication test-bed, and processing time & device utilisation details are reported. The performance of the proposed T–F measurement technique and designed digital architecture are critically investigated in terms of absolute error (AE), mean absolute error (MAE), root mean square error (RMSE), & correlation-coefficient (R). The obtained AE (over 80 time-bins), MAE, RMSE & R (over 10 frame-bins) is $\pm$1.24E−4, 0.076, 0.23, & 0.73, respectively. The obtained bit error rate (BER) (over the 5000 frame-bins) is 3.33E−4.