Organizations: Department of Computer Science, Rice University, Houston, United States · Department of Electrical and Computer Engineering, Duke University, Durham, United States
Automatic modulation classification (AMC) underpins spectrum sensing and non-cooperative signal identification in cognitive radio, yet the two standard CPU-feasible approaches to it, a compact convolutional neural network (CNN) on raw in-phase and quadrature (I/Q) samples and a classical higher-order-cumulant feature classifier, are rarely compared against each other conditioned on signal-to-noise ratio (SNR). We show that they are, in fact, complementary: the CNN is markedly more robust at low SNR, where hand-crafted statistics collapse toward chance, while the cumulant classifier is more accurate at high SNR, where higher-order statistics cleanly separate dense constellations. We exploit this complementarity with FuseAMC, a lightweight decision-fusion model that trains a logistic-regression meta-combiner over the concatenated class-posteriors of an approximately 198k-parameter residual CNN and a gradient-boosted-tree cumulant classifier, on a validation split disjoint from test. On RadioML 2016.10a (5 seeds, 44k-frame held-out test set), FuseAMC reaches 55.8% overall accuracy, improving on the best cumulant-feature baseline by 5.1 points and on the CNN alone by 3.2 points (paired-bootstrap 95% confidence-interval lower bounds of 5.0 and 3.1 points, both positive), while matching the best classical baseline in its own strong high-SNR regime and the CNN in its strong low-SNR regime. With about 0.2 million parameters, a model footprint of a few megabytes, and roughly 3 MFLOPs per frame, the whole pipeline trains and runs on one CPU core, targeting general-purpose and embedded processors without hardware acceleration.
Figures & tables
Method
Overall
Low
High
M-F1
Cumulant+SVM
0.507±0.001
0.241±0.001
0.843±0.001
0.525±0.001
Cumulant+RF
0.498±0.000
0.237±0.001
0.830±0.001
0.517±0.000
Cumulant+GBT
0.496±0.001
0.236±0.001
0.827±0.002
0.512±0.002
CNN (raw I/Q)
0.526±0.002
0.336±0.002
0.762±0.003
0.526±0.003
Avg. fusion
0.557±0.001
0.332±0.002
0.837±0.002
0.572±0.001
Fusion (ours)
0.558±0.001
0.337±0.002
0.835±0.001
0.575±0.001
TABLE I: Test accuracy by SNR regime (Low: SNR ≤ 0 dB; High: SNR ≥ 6 dB) and macro-F1 (M-F1) on RadioML 2016.10a (44k frames), mean ± std over 5 seeds. Best per column in bold.
Fig. 1: FuseAMC architecture: independent CNN (raw I/Q) and cumulant-feature+GBT streams each produce an 11-way posterior; a logistic-regression meta-combiner, trained on the validation split, fuses the concatenated posteriors into the final prediction.
Fig. 2: Test accuracy vs. SNR for FuseAMC, the CNN alone, and the best cumulant-feature baseline; lines are the 5-seed mean, shaded bands are plus/minus one standard deviation . The CNN and cumulant curves cross over near 0 dB (the CNN is higher below it, the cumulant classifier higher above it), and FuseAMC follows the upper envelope of both across the range.
Comparison
Δ Overall Acc.
95% CI
Fusion − CNN stream
+0.032
[+0.031,+0.034]
Fusion − GBT stream
+0.062
[+0.061,+0.063]
Fusion − Avg. fusion
+0.002
[+0.001,+0.003]
TABLE II: Fusion improvement in overall accuracy over each of its two streams (CNN, GBT) and over the naive posterior-average control, with 95% paired-bootstrap CIs over 5 seeds. Removing either stream, or replacing the learned combiner with a fixed average, strictly reduces accuracy.
Fusion rule
Overall acc.
Macro-F1
CNN stream
0.527±0.002
0.527±0.004
GBT stream
0.496±0.001
0.512±0.002
Posterior averaging
0.556±0.001
0.572±0.002
Weighted averaging
0.557±0.001
0.571±0.002
Temp.-calibrated avg.
0.556±0.001
0.572±0.002
SNR hard switch, oracle
0.551±0.002
0.562±0.003
TABLE III: Decision-fusion rules over the same frozen per-seed CNN and GBT posteriors, mean ± std over 5 seeds. Weighted averaging tunes its weight on validation ( αˉ=0.57 ); the oracle switch picks the SNR threshold on validation from the true SNR label. Every soft rule is within 0.4 points overall. Best in bold.
Stream
Calibration
Overall
Low SNR
High SNR
CNN
none
1.6
6.6
7.3
temperature ( T=1.06 )
1.5
5.6
7.8
isotonic
1.1
5.7
6.5
GBT
none
1.2
5.9
7.6
temperature ( T=0.95 )
1.3
6.7
7.2
isotonic
1.2
5.6
6.9
TABLE IV: Expected calibration error (ECE, %, 15 bins) of each stream’s posteriors, raw and after validation-fit temperature or isotonic calibration, over 5 seeds. Aggregate ECE is low for both streams, but rises to 6–8% within each SNR regime; calibrating the streams barely moves the learned combiner.
Test-time shift
CNN
GBT
Fusion
Δ Fus.
clean (reference)
0.527
0.496
0.559
—
sample-rate offset +1%
0.450
0.460
0.482
−0.077
2-tap fading, ∣a∣=0.3
0.409
0.469
0.470
−0.089
added noise, −6 dBc
0.448
0.398
0.464
−0.095
added noise, −3 dBc
0.351
0.325
0.361
−0.197
carrier offset 0.005 /sample
0.364
0.312
0.351
−0.208
TABLE V: Robustness to test-time channel/hardware shift absent from training: overall accuracy (mean over 5 seeds) when the 44k-frame test set is perturbed; models are trained only on the clean RadioML 2016.10a split. Δ is the fusion change from its clean accuracy.
Automatic modulation classification identifies radio signal modulation schemes from received waveforms, often under noise, fading, and limited labeled data.
Automatic modulation classification (AMC) is a key component of cognitive wireless systems, enabling spectrum awareness and adaptive communication. Most deep learning approaches to AMC assume a fixed set of modulation classes and access to previously observed training data. In practice, modulation classes and channel conditions may evolve over time, while storage or data-retention constraints may preclude replay of raw baseband I/Q samples. We propose PRISM-AMC, a replay-free continual learning framework that mitigates catastrophic forgetting by preserving task-relevant subspaces of intermediate representations via gradient projection memory. We further analyze projected stochastic-gradient updates under an SNR-dependent gradient-noise model. PRISM-AMC is evaluated in both class-incremental and domain-incremental settings. Experiments on the MIGOU and DeepSig 2018.01A datasets show substantially reduced forgetting relative to sequential fine-tuning and elastic weight consolidation, while remaining competitive with or outperforming replay-based RSD-AMC in class-incremental settings, without storing raw I/Q samples.
Usman Akram, Matthew Graham, Haris Vikalo
Department of Electrical and Computer Engineering, University of Texas at Austin, Austin, TX, USA · Applied Research Laboratories, Austin, TX, USA
Automatic modulation classification research has largely emphasized representation accuracy, but a cognitive receiver must also decide when heterogeneous evidence justifies overriding a trusted default prediction. We study this post-inference problem through cross-fitted residual utility and a primary-preserving cognitive decision policy. A structured KAN-Fourier classifier supplies the default probability, while neural and non-neural candidates provide observable evidence. Candidate-specific residual utility is learned from train-split out-of-fold predictions, and a disjoint validation split freezes action thresholds, approved transitions, conditional routes, and a unified risk mask before held-out evaluation. On RMLA, RMLB, and HISAR, the complete system improves overall accuracy from 63.632% to 66.332%, 65.161% to 66.168%, and 77.769% to 79.867%, respectively. Controlled comparisons show that the isolated utility target does not uniformly dominate alternative out-of-fold meta-learners; the consistent gain comes from the complete evidence-and-action policy. Paired bootstrap and Holm-corrected McNemar analyses support the controlled gains. A frozen-policy stress test under carrier-frequency offset, I/Q imbalance, and synthetic Rayleigh/Rician fading yields positive gains in all 11 conditions, with every paired 95% confidence interval above zero.
Linzhuo Han, Zongyong Cui, Houbiao Li
School of Mathematical Sciences, University of Electronic Science and Technology of China, Chengdu, China · School of Information and Communication Engineering, University of Electronic Science and Technology of China, Chengdu, China
Automatic modulation classification (AMC) of received radio signals is prudent for further signal processing tasks such as communication monitoring, cognitive radio operation, and interference mitigation in the electromagnetic spectrum. Traditional methods often rely on handcrafted features and struggle under complex channel conditions, whereas deep learning (DL) architectures can learn discriminative representations directly from raw signals. In this work, we propose HFEMCNet, a novel compact hybrid frequency-enriched multi-channel network that jointly exploits spatiotemporal dependencies and frequency domain information derived via Fast Fourier Transform (FFT). HFEMCNet integrates raw IQ samples with hierarchical spectral features to produce a rich signal representation, which is processed by convolutional layers for spatial feature extraction and Long Short-Term Memory (LSTM) units for temporal modeling. Experimental evaluation on benchmark datasets RML2016.10a, RML2016.10b, and over-the-air RML2018.01a demonstrates that HFEMCNet significantly outperform contemporary state-of-theart DL models in classification accuracy while achieving a reduced parameter count, smaller memory footprint, and lower tail-latency, making it suitable for real-time deployment on resource-constrained platforms. These results highlight the effectiveness of combining hybrid architectures with frequencydomain enrichment for robust and efficient AMC.