Reduced Complexity Ordered Statistics Decoding of Linear Block Codes

Abstract

This paper proposes a reduced complexity ordered statistics decoding (OSD) algorithm for linear block codes. With the received information, several most reliable positions in the ordered reliability sequence are prioritized as the validation band (VB). The Gaussian elimination (GE) is then performed to generate the systematic generator matrix of the code. In the process of re-encoding, only the test messages that satisfy the validation rule would be used to generate the codeword candidates, resulting in a low decoding complexity. The decoding error probability upper bound of the proposed OSD is further analyzed to characterize its performance-complexity trade off. Our simulation results show that the proposed OSD can significantly reduce the decoding complexity with a negligible loss in the decoding performance.