Abstract
In recent years, quasi-cyclic low-density parity-check code (QC-LDPC) is popularly used in IEEE 802.11n and 802.16e standard. In decoder part, Sum- Product Algorithm was generally used for decoding. The current issue is fo- cused on the throughput. If the throughput increases, the convergence of decoding was also increased. Sequential scheduling is a general way to speed up the convergence. LBP is one of them. But sequential scheduling was ‾xed the decoding process. It does not optimize the decoding process. To limit the convergence, informed dynamic scheduling (IDS) was proposed for fast conver- gence. IDS was used the decoding information to decided the next decoding step. Residual belief propagation (RBP) and Node-wise RBP (NW-RBP) were proposed in dynamic decoding process. But they have their own advantages and disadvantages. In this thesis, we proposed QUE-RBP method to combine their advantages and exceed the performance of NW-RBP. In another topic, we combine IDS with unequal error protection (UEP). We want to change the decoding process of RBP to make di□erent UEP levels. In many tests, we successfully use IDS and limit the APP to make di□erent UEP levels.