A Real-Time Deep Learning OFDM Receiver

Machine learning in the physical layer of communication systems holds the potential to improve performance and simplify design methodology. Many algorithms have been proposed; however, the model complexity is often unfeasible for real-time deployment. The real-time processing capability of these systems has not been proven yet. In this work, we propose a novel, less complex, fully connected neural network to perform channel estimation and signal detection in an orthogonal frequency division multiplexing system. The memory requirement, which is often the bottleneck for fully connected neural networks, is reduced by ≈ 27 times by applying known compression techniques in a three-step training process. Extensive experiments were performed for pruning and quantizing the weights of the neural network detector. Additionally, Huffman encoding was used on the weights to further reduce memory requirements. Based on this approach, we propose the first field-programmable gate array based, real-time capable neural network accelerator, specifically designed to accelerate the orthogonal frequency division multiplexing detector workload. The accelerator is synthesized for a Xilinx RFSoC field-programmable gate array, uses small-batch processing to increase throughput, efficiently supports branching neural networks, and implements superscalar Huffman decoders.

An identification of the tolerable time-interleaved analog-to-digital converter timing mismatch level in high-speed orthogonal frequency division multiplexing systems

<span>High-speed Terahertz communication systems has recently employed orthogonal frequency division multiplexing approach as it provides high spectral efficiency and avoids inter-symbol interference caused by dispersive channels. Such high-speed systems require extremely high-sampling <br /> time-interleaved analog-to-digital converters at the receiver. However, timing mismatch of time-interleaved analog-to-digital converters significantly causes system performance degradation. In this paper, to avoid such performance degradation induced by timing mismatch, we theoretically determine maximum tolerable mismatch levels for orthogonal frequency division multiplexing communication systems. To obtain these levels, we first propose an analytical method to derive the bit error rate formula for quadrature and pulse amplitude modulations in Rayleigh fading channels, assuming binary reflected gray code (BRGC) mapping. Further, from the derived bit error rate (BER) expressions, we reveal a threshold of timing mismatch level for which error floors produced by the mismatch will be smaller than a given BER. Simulation results demonstrate that if we preserve mismatch level smaller than 25% of this obtained threshold, the BER performance degradation is smaller than 0.5 dB as compared to the case without timing mismatch.</span>

Spectral Emission Mask Shaping for OFDM in Cognitive Radios

Orthogonal frequency division multiplexing (OFDM) is characterized by spectral efficiency. It enables flexible and agile spectrum allocation. But still it lags as it suffers from spectral leakage in the form of large side lobes. It leads to inter-channel interference if not handled carefully.in proposed system spectral emission mask system is implemented to combat spectral leakage and ultimately avoiding adjacent channel interference. A spectral mask, also known as a channel mask or transmission mask is a mathematically-defined set of lines applied to the levels of radio (or optical) transmissions. The spectral mask is generally intended to reduce adjacent-channel interference by limiting excessive radiation at frequencies beyond the necessary bandwidth. The proposed system is implemented over MATLAB platform using script language.

Comparison of SPS and FA methods for Blind Carrier Frequency Offset Estimation in OFDM systems

In orthogonal frequency-division multiplexing (OFDM) systems, synchronization issues are of great importance since synchronization errors might destroy the orthogonality among all subcarriers and, therefore, introduce intercarrier interference (ICI) and intersymbol interference (ISI). Several schemes of frequency offset estimation in OFDM systems have been investigated. This paper compares performance and computational complexity of Smoothing Power Spectrum (SPS) and Frequency Analysis (FA) methods for blind carrier frequency offset (CFO) estimation in OFDM systems.

Generalized Frequency Division Multiplexing-Based Low-Power Underwater Acoustic Image Transceiver

In this paper, a low-power underwater acoustic (UWA) image transceiver based on generalized frequency division multiplexing (GFDM) modulation for underwater communication is proposed. The proposed transceiver integrates a low-density parity-check code error protection scheme, adaptive 4-quadrature amplitude modulation (QAM) and 16-QAM strategies, GFDM modulation, and a power assignment mechanism in an UWA image communication environment. The transmission bit error rates (BERs), the peak signal-to-noise ratios (PSNRs) of the received underwater images, and the power-saving ratio (PSR) of the proposed transceiver obtained using 4-QAM and 16-QAM, with perfect channel estimation, and channel estimation errors (CEEs) of 5%, 10%, and 20% were simulated. The PSNR of the received underwater image is 44.46 dB when using 4-QAM with a CEE of 10%. In contrast, PSNR is 48.79 dB when using 16-QAM with a CEE of 10%. When BER is 10−4, the received UW images have high PSNR values and high resolutions, indicating that the proposed transceiver is suitable for underwater image sensor signal transmission.