Skip to main content
Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

< Back to Article

Table 1.

Comparison of existing schemes.

More »

Table 1 Expand

Fig 1.

Deep learning-based AFDM system architecture.

More »

Fig 1 Expand

Fig 2.

Probability density function of the AFDM signal amplitude.

More »

Fig 2 Expand

Fig 3.

Impact of the conventional companding transform method on the signal.

More »

Fig 3 Expand

Fig 4.

PolyNet-Volterra Network Architecture.

More »

Fig 4 Expand

Table 2.

Simulation parameter configuration.

More »

Table 2 Expand

Fig 5.

Convergence process during training under a fixed SNR (20 dB) environment.

More »

Fig 5 Expand

Fig 6.

Convergence process during training under a dynamic SNR (5 dB to 25 dB) environment.

More »

Fig 6 Expand

Table 3.

Benchmark model configurations and comparison rationale.

More »

Table 3 Expand

Table 4.

Computational complexity and performance scores of different models ().

More »

Table 4 Expand

Table 5.

BER of each scheme under different SNRs (0-25 dB).

More »

Table 5 Expand

Fig 7.

Bit Error Rate (BER) performance comparison under different SNRs (0-25 dB).

More »

Fig 7 Expand

Table 6.

Feature-subset models.

More »

Table 6 Expand

Fig 8.

Comparison of reconstruction MSE in the masked region across different feature combinations.

More »

Fig 8 Expand

Fig 9.

Histogram of parameter efficiency at 15 dB.

More »

Fig 9 Expand

Fig 10.

Heat map analysis of the correlation between various input features and nonlinear distortion errors at a 10 dB signal-to-noise ratio.

More »

Fig 10 Expand

Fig 11.

Heat map analysis of the correlation between various input features and nonlinear distortion errors at a 15 dB signal-to-noise ratio.

More »

Fig 11 Expand

Fig 12.

Heat map analysis of the correlation between various input features and nonlinear distortion errors at a 20 dB signal-to-noise ratio.

More »

Fig 12 Expand

Fig 13.

Scatter plot of the reconstruction error in the masked region versus the original signal power (SNR = 15 dB).

More »

Fig 13 Expand

Fig 14.

Gradient magnitude and relative contribution of the input features.

More »

Fig 14 Expand

Fig 15.

Feature weight distribution of the first-layer hidden neurons.

More »

Fig 15 Expand

Fig 16.

Mean square error of various schemes at the masked positions.

More »

Fig 16 Expand

Table 7.

Repeated independent Monte Carlo variability at 20 dB.

More »

Table 7 Expand

Table 8.

Uncertainty quantification with BER and masked-region MSE confidence intervals.

More »

Table 8 Expand

Table 9.

Formal significance tests against PolyNet-Volterra at 20 dB.

More »

Table 9 Expand

Fig 17.

Extended statistical validation at 20 dB, including repeated-run BER distributions, Wilson confidence intervals, and aggregate significance levels.

More »

Fig 17 Expand

Table 10.

Sensitivity to the masking threshold at 20 dB.

More »

Table 10 Expand

Fig 18.

Sensitivity of PAPR reduction and reconstruction performance to the masking threshold at 20 dB.

More »

Fig 18 Expand

Table 11.

High-order modulation experiment with modulation-specific PolyNet-Volterra training.

More »

Table 11 Expand

Fig 19.

High-order modulation performance after modulation-specific PolyNet-Volterra training.

More »

Fig 19 Expand

Table 12.

Doppler robustness of PolyNet-Volterra at 20 dB.

More »

Table 12 Expand

Fig 20.

Doppler robustness of PolyNet-Volterra under increased maximum Doppler index.

More »

Fig 20 Expand

Table 13.

Complexity and RTX 3090 GPU inference latency of neural reconstruction models.

More »

Table 13 Expand

Table 14.

Measured RTX 3090 GPU training time for the 150-epoch training schedule.

More »

Table 14 Expand

Table 15.

Full-frame input ablation at 20 dB.

More »

Table 15 Expand

Fig 21.

Full-frame input ablation comparing full-frame and masked-only neural reconstruction.

More »

Fig 21 Expand