Recovery of unknown constraint length and generator polynomials for linear convolutional encoders

There are cases where the parameters of a forward error correcting encoder, the constraint length K and the generator polynomials, are unknown and must be obtained before the message data can be recovered. An algorithm is presented which, through observation and manipulation of the encoder data, composes a synthetic impulse response and from this sequence performs encoder parameter recovery. With these parameters in hand, selection of a suitable (Viterbi) decoder is straightforward, as is recovery of message data. The class of encoder examined is feedback-free, rate 1/2, linear, convolutional encoders, and for simplicity the algorithm is presented assuming noise-free data. The algorithm is then generalized to accommodate rate m/n codes.