Skip to Main Content
A synchronization system for use with noisy digital signals and a theoretical analysis of its operation is given. The system uses bit-by-bit correlation of the random bit stream input to determine synchronization error. The contention is that any periodically varying signal may be correlated period-by-period to find relative period synchronization. No knowledge of the information content of the signal is required. As conceived, this system is applicable to any random function which has periodicity. Since it uses correlation, it will enable synchronization in relatively low signal-to-noise ratio systems. The paper is a comparative analysis of periodic crosscorrelation detection vs. periodic integration of a diphase (Manchester) coded signal of fixed period and random digits.