ANALYTIC DESIGN OF CONVOLUTIONAL ENCODERS.
Date of Original Version
Summary form only given. Binary convolutional encoders are classified by their constraint length L, the number of input bits k, and the number of output bits n with the coder's details contained in an L multiplied by n binary interconnection matrix H (with rows h//i ). Performance is typically measured through free distance. An analytuc approach is presented to the maximization of free distance which a priori fixes only L and k. The technique first computes the free distance as a function of the L unknown rows of H and then selects the h//i to maximize this function. It involves specifying the encoder's state diagram (with branch weights being functions of the //i ) and computing the encoders transfer function through Mason's gain formula, keeping only those terms relevant to the free distance. An example is given.
Swaszek, Peter F.. "ANALYTIC DESIGN OF CONVOLUTIONAL ENCODERS.." , (1986): 82. https://digitalcommons.uri.edu/ele_facpubs/1115