In this paper, we present a comprehensive survey and detailed comparison of techniques that have been applied to the problem of identifying the type of modulation contained within received wireless signals. Known as automatic modulation classification (AMC), the problem has been studied for many decades. AMC plays a significant role in both military and civilian scenarios and is the main step in smart receivers. Especially with the development of software-defined radios and automatic communication systems, IoT technology and the spread of 5G technology bring the number of spectrum-using equipment explosion, making the problem of scarce spectrum resources more prominent. Although AMC techniques can be optimized from the classifier's point of view, signal pre-processing also plays a critical role. Relevant data representation approaches include time-frequency analysis, cyclostationary transforms, and hybrid techniques. We provide a taxonomy of common approaches based on order and dimensionality along with an overall analysis of signal pre-processing algorithms for AMC. Furthermore, we reproduce the major existing schemes under uniform conditions, allowing an objective comparison among different methodologies. Finally, we create an open-source Python library to simulate these techniques so the results in this paper are reproducible for future research.

The spectral correlation density (SCD) is an important tool in cyclostationary signal detection and classification. Even using efficient techniques based on the fast Fourier transform (FFT), real-time implementations are challenging because of the high computational complexity. A key dimension for computational optimization lies in minimizing the wordlength employed. In this paper, we analyze the relationship between wordlength and signal-to-quantization noise in fixed-point implementations of the SCD function. A canonical SCD estimation algorithm, the FFT accumulation method (FAM) using fixed-point arithmetic is studied. We derive closed-form expressions for SQNR and compare them at wordlengths ranging from 14 to 26 bits. The differences between the calculated SQNR and bit-exact simulations are less than 1 dB. Furthermore, an HLS-based FPGA design is implemented on a Xilinx Zynq UltraScale+ XCZU28DR-2FFVG1517E RFSoC. Using less than 25% of the logic fabric on the device, it consumes 7.7 W total on-chip power and has a power efficiency of 12.4 GOPS/W, which is an order of magnitude improvement over an Nvidia Tesla K40 graphics processing unit (GPU) implementation. In terms of throughput, it achieves 50 MS/sec, which is a speedup of 1.6 over a recent optimized FPGA implementation.

The spectral correlation density (SCD) function is the time-averaged correlation of two spectral components, used for analysing periodic signals with time-varying spectral content. Although the analysis is extremely powerful, it has not been widely adopted in real-time applications due to its high computational complexity. In this paper, we present an efficient FPGA implementation of the FFT accumulation method (FAM) for estimating the SCD function and its alpha profile. The implementation uses a linear systolic array with a bi-directional datapath consisting of DSP-based processing elements (PEs) with a dedicated instruction schedule, achieving a PE utilization of 88.2%. The 128-PE implementation achieves a clock frequency in excess of 530 MHz and consumes 151K LUTs, 151K FFs, 264 BRAMs, 4 URAMs and 1054 DSPs, which is less than 36% of the logic fabric on a Zynq UltraScale+ XCZU28DR-2FFVG1517E RFSoC device. It has a modest 12.5W power consumption and an energy efficiency of 4832 MOPS/W which is 20.6 × better than the published state-of-the-art GPU implementation. In terms of throughput, it achieves 15340 windows/s (15340 windows/s  × 2048 samples/window = 31.4 MS/s), which is a 4.65 × improvement compared to the above-mentioned GPU implementation and 807 × compared to an existing hybrid FPGA-GPU implementation.