Fwd: decode problem in benchmark--NCO?

For BPSK, not using differential, is there any way to keep track of the
rotating so as to understand exactly what happened to each bit? Only
understanding what happens to each bit in a single series of data, I can
work on how to joint them correctly.

Other than differential decoding, another way to resolve phase-ambiguity
to send Unique-word (header) in the start. This unique-word will be
complemented in BPSK if ther is phase-ambiguity in the received data.