Using ARM for signal processing

I don’t have access to a signal processor in my design. However, I have to process multiple signals on a single (wireless) bus. The way they are distinguished is via frequency and phase offset. There are three separate frequencies that must be distinguished from each other.

I would like to perform a series of correlations but afraid that would quickly overload the processor (it also has to run an OS and numerous other tasks). Is there a simpler method to distinguish several slow frequencies? The signals are all at around 10 Hz.

Thanks!

your goal isn’t very clear, but here goes…

if you have several 10Hz type signals…

Let’s assume they’re analog signals.

You’d sample them at 20Hz or more. Say 100Hz.

A chip with a multi-channel A/D would do. Sample the signals in succession, if that’s OK. The A/D rate might be 200KHz max sample rate at max.

Faster with a separate A/D.

Then what? FFT/IFFT/DFFT?

Just an amplitude vs. time histogram?

Then what?

Goertzel algorithm? But I think the point at which a FFT is better than multiple Goertzels is around 3.

Are the signals simultaneous, or is this more like an FSK decoder? What’s the bandwidth of each individual signal?