Sparse Parity-Check Matrix Encoding for MIMO OFDM Error Bursts

The patent segments the data stream into multiple parallel streams and applies different encoding schemes to different segments. Specifically, it uses convolutional encoding for some streams and LDPC encoding for others, then interleaves them. This segmentation allows the system to prevent long sequences of noisy bits by distributing errors across different encoded segments rather than having them concentrated in one location.

The patent introduces an interleaver as an intermediary component between the encoding stage and the transmission stage. This interleaver reorders the encoded bits from multiple streams, acting as a mediator that distributes adjacent noisy bits across different code words and streams. This intermediary operation breaks up long sequences of noisy bits into shorter, more manageable error patterns that can be corrected by the decoding process.

The patent merges multiple encoding techniques (convolutional encoding and LDPC encoding) into a hybrid system. By combining the strengths of both approaches - the simplicity and robustness of convolutional codes with the high performance of LDPC codes - the system achieves superior error correction capability while managing complexity through structured integration. The merged approach allows optional use of LDPC while maintaining compatibility with simpler convolutional encoding paths.

The patent implements a dynamic encoding system where the choice between convolutional encoding and LDPC encoding can be adjusted based on channel conditions and performance requirements. The system dynamically selects encoding parameters such as code rate, block length, and encoding type to optimize the trade-off between error correction performance and computational complexity for different operating scenarios.