Synchronisation techniques for a CCMA system using neural nets

A new synchronisation technique for a T out of N-user multiple access adder channel (MAC) with binary inputs is described. The application of artifici al neural networks at the decoder is considered as a fast and robust techni que for filtering, sorting, shifting and decision malting upon the incoming MAC data. An efficient decoding technique is introduced for the extraction of block synchronisation information, thus helping to reconstruct the orig inal signals accurately.