In communications systems, there must be a method to identify the beginning of a message at the receiving end, to know when the data stream of a specified structure begins. This is conventionally done by inserting attached sync markers (ASMs) between codewords, and looking for these ASMs in the transmitted data. Instead, a means has been developed to detect the positions of codewords directly, using the decoder to partially process a decoder attempt at each candidate synchronization position. A short decoding operation can successfully synchronize codewords in about two seconds in software.
This method is a variation of the brute force approach in which decoding is halted prematurely. At the correct offset, the messages passed in the decoding algorithm begin to converge in a fundamentally different way than they do when the offset is incorrect. This difference can be exploited by forming an appropriate metric that discriminates between the correct offset and the incorrect offsets.
A brute force way to synchronize frames is to buffer two frame-lengths of symbols — a length sufficient to guarantee capture of at least one full frame — and attempt decoding at each possible offset until an offset is found for which decoding is successful. This decode-at-all-offsets approach was used in the Mars Laser Communications Demonstration (MLCD), for example, and works well if the decoder is many times faster than the data rate of the link. For CCSDS LDPC (Consultative Committee for Space Data Systems Low-Density Parity-Check) codes, the decoder would need to operate four to five orders of magnitude faster than the data rate in order for it to be able to acquire the correct frame offset without dropping or buffering additional codewords during the synchronization process.
This method is computationally more complex and doesn’t perform as well as frame synchronizers that utilize an ASM; nevertheless, the new synchronizer acquires frame synchronization in about two seconds when using a 600-kbps software decoder, and would take about 15 ms on prototype hardware. It also eliminates the need for the ASMs, which is an attractive feature for short uplink codes whose coding gain would be diminished by the overheard of ASM bits.
This work was done by Jon Hamkins of Caltech for NASA’s Jet Propulsion Laboratory.
This Brief includes a Technical Support Package (TSP).
Frame Synchronization Without Attached Sync Markers
(reference NPO47920) is currently available for download from the TSP library.
Don't have an account? Sign up here.