5G-NR Signal Decoding on 5G-NR
0.0005FLOPs (GFLOPs)PCSI-Rx
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| PCSI-RxType=Traditional 5G-NR Receiver, Params (K)=02025.12 | 0.0005 | 2.5 | |
| LS-RxType=Traditional 5G-NR Receiver, Params (K)=02025.12 | 0.0006 | 2.71 | |
| NeuromorphicRxArchitecture=SNN, Depth=5 ResNets, Params (K)=190.342025.12 | 1.15 | 1,003.33 | |
| NeuralRxArchitecture=ANN, Depth=5 ResNets, Params (K)=190.342025.12 | 1.36 | 6,671.79 | |
| NeuromorphicRxArchitecture=SNN, T=2, Quantization=8-bit, Params (K)=262.532025.12 | 1.46 | 260.62 | |
| NeuromorphicRxArchitecture=SNN, Depth=7 ResNets, Params (K)=262.532025.12 | 1.51 | 1,207.2 | |
| NeuromorphicRxArchitecture=SNN, T=2, Params (K)=262.532025.12 | 1.51 | 1,207.2 | |
| NeuromorphicRxArchitecture=SNN, Depth=9 ResNets, Params (K)=334.722025.12 | 1.77 | 1,326.26 | |
| NeuralRxArchitecture=ANN, Depth=7 ResNets, Params (K)=262.532025.12 | 1.88 | 9,202.91 | |
| NeuralRx (Ours baseline)Architecture=ANN, Params (K)=262.532025.12 | 1.88 | 9,202.91 | |
| NeuromorphicRxArchitecture=SNN, Depth=11 ResNets, Params (K)=406.922025.12 | 2.07 | 1,474.09 | |
| NeuromorphicRxArchitecture=SNN, T=10, Params (K)=262.532025.12 | 2.22 | 5,112.72 | |
| NeuralRxSource=Ref [7], Params (K)=185.992025.12 | 2.38 | 11,667.72 | |
| NeuralRxArchitecture=ANN, Depth=9 ResNets, Params (K)=334.722025.12 | 2.39 | 11,734.03 | |
| NeuralRxArchitecture=ANN, Depth=11 ResNets, Params (K)=406.922025.12 | 2.91 | 14,265.15 | |
| NeuralRxSource=Ref [4], Params (K)=604.232025.12 | 4.99 | 24,480.66 | |
| LMMSE-RxType=Traditional 5G-NR Receiver, Params (K)=02025.12 | 3,410.66 | 16,712.24 |