Listing 1. CNN structure used for signal classification.

1. sig_inp = Input(shape=(4990,1)) # input signal

2. conv1 = Conv1D(16, 20, activation=“relu”)(sig_inp)

3. pool1 = MaxPooling1D(pool_size=3)(conv1)

4. conv2 = Conv1D(16, 10, activation=“relu”)(pool1)

5. pool2 = MaxPooling1D(pool_size=3)(conv2)

6. conv3 = Conv1D(16, 10, activation=“relu”)(pool2)

7. pool3 = MaxPooling1D(pool_size=3)(conv3)

8. flat = Flatten()(pool3)

9. dense1 = Dense(64, activation=“relu”)(flat)

10. output = Dense(4, activation=“softmax”)(dense1) # output class