Comparison of Neural Network Architectures for Spectrum Sensing
This work addresses the problem of selecting optimal neural network architectures for spectrum sensing in communication systems, but it is incremental as it compares existing methods without introducing new ones.
The paper compared four neural network architectures (FC, CNN, RNN, BiRNN) for spectrum sensing to determine the best fit for classifying communication signals, finding that CNN, RNN, and BiRNN achieved similar performance with abundant resources, while FC performed worse except under strict computational limits.
Different neural network (NN) architectures have different advantages. Convolutional neural networks (CNNs) achieved enormous success in computer vision, while recurrent neural networks (RNNs) gained popularity in speech recognition. It is not known which type of NN architecture is the best fit for classification of communication signals. In this work, we compare the behavior of fully-connected NN (FC), CNN, RNN, and bi-directional RNN (BiRNN) in a spectrum sensing task. The four NN architectures are compared on their detection performance, requirement of training data, computational complexity, and memory requirement. Given abundant training data and computational and memory resources, CNN, RNN, and BiRNN are shown to achieve similar performance. The performance of FC is worse than that of the other three types, except in the case where computational complexity is stringently limited.