Steps:
1. A/D conversion, store result in buffer
2. Stop A/D when buffer is full
3. Perform FFT on signal buffers
4. Check for common motion on all results (device moving), and set FFT bin value to 0
5. Detect signals above the noise threshold for each of the FFT resulting bin
6. For each of the signals that are above noise create an entry.
7. Extrapolate moving object angle by interpolating a polynomial between two adjacent entries.
	Explanation:	When an moving object is being detected by two radar modules, both will detect it.
					To determinate the angle of the object some interpolation must be performed,
					one option is to try to fit a parabola between the two values and find the local extreme (todo: sketch)
8. Display results on display

- All of the steps go in a pipeline, step 1 is ran by a periodic interrupt.
- Once per second FFT is initialised for each set of new signal data.
- Max frequency is 100Hz, sample rate 200Hz, sample time 0.5 seconds
- Depending oh how much RAM is left multiple signals can be processed simultaneously
- If FFT turns out to be too difficult to implement the device will still be able to operate, but not while moving
- Accelerometer can be used to help determinate if device is moving
- Magnetometer can be used as compas on display
- Refresh cycle will be relatively slow, something like in the movie/game
- 18F4525 can be used for more RAM
- Sound effects are generated by the SAA1099
- FMCW can be achieved by using the D/A converter