This repository is a sandbox for the BPSK demodulation quiz.
Results can be found in the "results" directory.
- I was a little uncertain about the timing assumption mentioned in the problem statement. I mentioned methods for timing recovering in the code but assumed timing was found.
- There is definitely some tuning that could be done to improve convergence rates. The provided "time" for convergence was 1 second. I interpretted this as meaning up to 1e6 symbols and must not perform any computationally heavy tasks that would exceed 1 second.
- I had never actually implmented a PLL before (it was fun!) but I imagine that with some more work, convergence could be hastened with higher order methods and that stability could be improved by incorporating more sample memory into iterations.
- Some of the image saving code is sort of ugly but I wanted to ensure readers should easily run.
- Using pilots would provide ambiguity resolution.
- Using a differential bpsk would ease implementation as phase ambiguity does not impede detection.