Deep Learning-based Carrier Frequency Offset Estimation with One-Bit ADCs

With an increase in defined carrier frequency for the 5G new radio system, the need for synchronization also increases. The transceiver loses synchronization due to the occurrence of timing and carrier frequency offsets. Carrier frequency offsets often occur due to mismatch between transmitter and receiver oscillator frequency as well as the occurrence of doppler shifts due to transmitter/receiver movements. When frequency offsets exceed subcarrier spacing, integer frequency offset occurs that results in performance loss due to subcarrier indices shifts. Conventional approach i.e. maximum likelihood and sequential method is already employed to estimate integer frequency offset and to detect sector id. In this paper, the deep learning-based method is demonstrated to estimate the integer frequency offset and sector id detection. The neural network containing multiple convolution layers with activation layers is used to find the optimum received signal. Then, by calculating the number of cyclic shifts in the optimum received signal, the integer frequency offset is estimated. Using the corrected optimum received signal, the primary synchronization signal is also detected that gives sector id. This proposed estimator is tested for different profiles of tapped delay line models with different desired delay spread and compared with conventional methods i.e. maximum likelihood estimation method and sequential estimation method. Simulation results show that the proposed Neural Network based estimator outperforms in all delay profiles.

show abstract

“…In [17], CFO of a complex sinusoid is estimated using Deep Learning (DL) architectures. The achieved estimation range of CFO is [0.2,0.25].…”

Section: Deep Learning Based Carrier Frequency Offsets (Cfos) Estimation Methodsmentioning

confidence: 99%

A Neural Network based Integer Frequency Offset Estimation and PSS Detection in 5G NR Systems

Patel

Warhade²

2021

IJIES

View full text Add to dashboard Cite

show abstract

“…Authors in [21] address the problem of CFO in the uplink of the OFDM access (OFDMA) system, where DL is used to suboptimally estimate CFOs corresponding to different users. The DL-based CFO for the received signals after a low resolution analog-to-digital conversion in emerging mmWave multiple-input multiple-output (MIMO) systems is investigated in [22], demonstrating improved performance as compared to the conventional methods. For OFDM-based unmanned aerial vehicle communications, DL methods for CFO are proposed in [23].…”

Section: A Related Work and Paper Contributionsmentioning

confidence: 99%

Deep Learning-Based Packet Detection and Carrier Frequency Offset Estimation in IEEE 802.11ah

Ninković

Valka²,

Dumic³

et al. 2021

IEEE Access

View full text Add to dashboard Cite

Wi-Fi systems based on the IEEE 802.11 standards are the most popular wireless interfaces that use Listen Before Talk (LBT) method for channel access. The distinctive feature of a majority of LBTbased systems is that the transmitters use preambles that precede the data to allow the receivers to perform packet detection and carrier frequency offset (CFO) estimation. Preambles usually contain repetitions of training symbols with good correlation properties, while conventional digital receivers apply correlationbased methods for both packet detection and CFO estimation. However, in recent years, data-based machine learning methods are disrupting physical layer research. Promising results have been presented, in particular, in the domain of deep learning (DL)-based channel estimation. In this paper, we present a performance and complexity analysis of packet detection and CFO estimation using both the conventional and the DL-based approaches. The goal of the study is to investigate under which conditions the performance of the DLbased methods approach or even surpass the conventional methods, but also, under which conditions their performance is inferior. Focusing on the emerging IEEE 802.11ah standard, our investigation uses both the standard-based simulated environment, and a real-world testbed based on Software Defined Radios.

show abstract

“…To benchmark our multi-sinusoid estimator, we employ the Periodogram, because it has been shown to produce better estimates for low resolution sinusoidal data than eigendecomposition and dithering methods [12], [29]. The Periodogram is calculated from the scaled-and-squared, zero-padded Fast Fourier Transform (FFT) with N 0 = 2 16 to ensure that grid resolution is not a limiting factor.…”

Section: Benchmarksmentioning

confidence: 99%

“…The use cases for such algorithms are broad, with varying compute constraints (server, cloud, deployed), but our algorithm is primarily intended for edge computer signal processing, so we also consider important characteristics such as memory, training sample size, and execution time in our investigation. From our past work, we found convolutional neural networks were sufficiently powerful and fast for near real-time estimation of a single sinusoid [12]. As a result, our networks rely heavily on convolutional layers for parameter sharing, dimensionality reduction, and information parsing.…”

Section: Introductionmentioning

confidence: 97%

“…These methods rely on knowledge of the amplitude of the signal, or the ability to estimate by iterative grid search. Unlike the unquantized case, non-parametric methods, such as the Periodogram, are often the best estimators with extremely low-resolution sinusoid estimation [12], [29]. All of these techniques (non-parametric estimators [21], correlation-based estimators [28], dither [29] and time-varying thresholds [30]) do not make use of the quantization noise correlation [11].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

SignalNet: A Low Resolution Sinusoid Decomposition and Estimation Network

Dreifuerst,

Heath

2021

Preprint

Self Cite

View full text Add to dashboard Cite

The detection and estimation of sinusoids is a fundamental signal processing task for many applications related to sensing and communications. While algorithms have been proposed for this setting, quantization is a critical, but often ignored modeling effect. In wireless communications, estimation with low resolution data converters is relevant for reduced power consumption in wideband receivers. Similarly, low resolution sampling in imaging and spectrum sensing allows for efficient data collection. In this work, we propose SignalNet, a neural network architecture that detects the number of sinusoids and estimates their parameters from quantized in-phase and quadrature samples. We incorporate signal reconstruction internally as domain knowledge within the network to enhance learning and surpass traditional algorithms in mean squared error and Chamfer error. We introduce a worst-case learning threshold for comparing the results of our network relative to the underlying data distributions. This threshold provides insight into why neural networks tend to outperform traditional methods and into the learned relationships between the input and output distributions. In simulation, we find that our algorithm is always able to surpass the threshold for three-bit data but often cannot exceed the threshold for one-bit data. We use the learning threshold to explain, in the one-bit case, how our estimators learn to minimize the distributional loss, rather than learn features from the data.

show abstract

Deep Learning-based Carrier Frequency Offset Estimation with One-Bit ADCs

Cited by 24 publications

References 11 publications

A Neural Network based Integer Frequency Offset Estimation and PSS Detection in 5G NR Systems

A Neural Network based Integer Frequency Offset Estimation and PSS Detection in 5G NR Systems

Deep Learning-Based Packet Detection and Carrier Frequency Offset Estimation in IEEE 802.11ah

SignalNet: A Low Resolution Sinusoid Decomposition and Estimation Network

Contact Info

Product

Resources

About