Computationally efficient implementation for incremental redundancy within wireless networks

Hybrid ARQ (HARQ) is widely used nowadays for error control within diverse wireless networks. HARQ deploys error detection mechanisms to corrected code sequences and enforces successive retransmissions and decoding until the corrected code sequence is presumed to be decoded correctly or a maximum number of retransmissions are reached. Incremental redundancy is a HARQ technique where additional transmissions associated with error correction failures are comprised only of incremental coding bits that increase the robustness of the channel coding. Nevertheless, the throughput and BER advantages of incremental redundancy should not be treated in isolation from the computational cost associated with repetitive decoding of the same information data. In this paper, simple and computationally efficient approaches are proposed for the implementation of incremental redundancy based on rate compatible punctured convolutional codes under hard-decision decoding. The proposed approaches identify code sequences over which the cost-metric remains constant. These sequences are assumed reliable and their associated computations are eliminated whenever decoding should be repeated at lower rates. The proposed approaches are evaluated from the perspective of computations, memory requirements. The corresponding effects on bit error rate and throughput performance are reported. The computational gains (up to 21%) of the proposed approaches would be particularly beneficial in systems where energy resources are scarce such as wireless sensor networks.