Classification of digital signals using neural networks?

Question

Mathematics

There are two types of digital signals (it is not known in advance which signal belongs to which type). The task is to classify them in real time. Is such a problem solvable for neural networks?

Answer 1

Why is there a neural network? This classical problem is solved either by analyzing the components of the spectrum (Fourier/wavelet transform) or by cross-correlation function.

Answer 2

In any case, neural networks require a function for evaluating the result, whether the answer from the neural network is correct or not. That is, if it is not possible to evaluate the result (both the input data and the result in the form of a signal type are not known), then it will not work to train the neural network.
but unsupervised learning, a problem to be solved, especially for clustering, such clustering will be simply, for example, by the "proximity" of clusters to each other, relative to the parameters (without any connection with the real result)