Blind Recovery of Convolutional Codes over Noisy Channels

In this paper we study blind recovery methods for identifying the parameters of a convolutional code from an intercepted bitstream. It is assumed that the interceptor has no prior knowledge about the encoder and the recorded bitstream is noisy. The goal is to obtain the following parameters of convolutional code (i) number of inputs (ii) number of outputs (iii) constraint length of the convolutional encoder. Given these parameters one can regenerate the encoder and fully decode the intercepted bitstream. To evaluate our proposed method, we model the effect of noise on bitstream by errors in binary symmetric channel and burst format. Our simulation results show that the proposed method can accurately extract the parameters of convolutional codes in various scenarios.