Low Density Parity Check (LDPC) codes are considered in many future communication systems for error correction coding. Optimal decoding of LDPC codes is usually too costly to be done in practice. For this reason, sub-optimal algorithms are used. A state-of-the-art algorithm for decoding of LDPC codes is called belief propagation (BP). For short LDPC codes and for codes with an implementation efficient structure, the performance of this algorithm can be far from optimum.
We present a graphical model for representing the decoding problem, called configuration graph. We show the construction of a configuration graph and describe how the decoding problem can be represented as maximum weighted vertex problem (VP) on a configuration graph. We describe decoding approaches utilizing this representation and show the improvements in terms of decoding performance as well as the complexity/performance trade-offs possible with these algorithms.
Bitte loggen Sie sich ein, um Zugang zu diesem Inhalt zu erhalten