I have some signals and I want to classify them into different types. Signals are all sine shape signals, some might have glitches. I know I could classify them base on frequency or amplitude, but is there other ways to classify signals? Thank you.
10
-
$\begingroup$ One can encode signals into waves in only 3 ways, amplitude of the wave, frequency of the wave, or phase of the wave (or combinations of them). If phase of the wave is not used in the encoding, for example, it is not necessary to analyze this phase. Ultimately, a signal is good when the original message is decodable, or it is bad when the original message is garbled. There are standard metrics like bits error rate (ber) which compares the number of bits we get right versus the number we get wrong, which can be used to evaluate the system's signal performance. $\endgroup$– JamesCommented Dec 2, 2022 at 15:29
-
$\begingroup$ It will depend on the context you want to use. For example, you might classify on some characteristic that relates to information you are trying to transmit. $\endgroup$– Boba FitCommented Dec 2, 2022 at 15:32
-
$\begingroup$ Please clarify your specific problem or provide additional details to highlight exactly what you need. As it's currently written, it's hard to tell exactly what you're asking. $\endgroup$– Community BotCommented Dec 2, 2022 at 15:38
-
$\begingroup$ A data-driven approach would be to use something like k-means clustering or principal component analysis to see if your data naturally breaks into distinguishable classes. However that would be more of a question relevant at the cross-validated stack exchange. The physics version of this question would be something like, how can you understand the processes that produce glitches and how those processes produce the glitch morphology you see. $\endgroup$– AndrewCommented Dec 2, 2022 at 15:39
-
$\begingroup$ @James there are many more ways than only three, just decompose a time function into an arbitrary complete orthogonal set of functions and use the coefficients as information carriers: no phase, no amplitude, no frequency. Those three are the simplest waveform coding schemes and historically the first because of the available narrow band passive filters/antennas and modulating/demodulating electronics (AM/FM/PM). $\endgroup$– hyportnexCommented Dec 2, 2022 at 17:06
|
Show 5 more comments