Polar subcodes for encoding and blind decoding of variable-sized data blocks

Abstract:
A scheme for encoding variable-sized data blocks and their low-complexity blind decoding, i.e. decoding without the information about an actual size of the payload, is presented. This scheme relies on polar subcodes and includes the code design rule as well as the idea of the decoding algorithm. A code construction is suggested, which modifies the previously constructed family of polar subcodes. As an example of low-complexity blind decoding algorithm, a generalization of the sequential decoder is presented, which is able to jointly detect the payload size and correct errors in the received noisy sequence. Simulation results are provided and demonstrate the significant complexity reduction in comparison to the case of multiple decoding attempts and the performance close to the independent decoding of each code.