Goodness Analysis of Generator Polynomial for Convolution Code with Varying Constraint Length

Abstract: Convolutional codes play a vital role in reducing the error produced due to dispersive effect, fading, attenuation, interference and noise during transmission and reception of data. With a proper design, convolutional encoder and Viterbi decoder can jointly provide an acceptable performance with feasible decoding complexity. In this combination, a tradeoff on the error performance and decoding complexity resides on the choice of the constraint length. However, the increase in code constraint length leads to exponential increase in the computational complexity of the Viterbi decoder. In this paper, a simulink model with convolution encoder and viterbi decoder for � code rate is designed with varying constraint length from 3 to 10. The bit error rate performance has been analyzed for different generator polynomials and their goodness is examined with the motive of proposing the best generator polynomial at different SNR values.

Keywords: Convolutional codes, BPSK, AWGN, Viterbi Algorithm, Generator polynomial.