A Family of Quasi-Cyclic LDPC Codes for Deep Space Communications
- Paper number
IAC-06-B3.2.10
- Author
Dr. He Shanbao, China Academy of Space Technology (CAST), China
- Year
2006
- Abstract
Generally, the distance between the earth and spacecraft is enormous for deep space exploration and communications. So power efficiency is a particularly important metric for deep-space applications because the typical received signal power is in inverse proportion to the distance. Low-density parity-check (LDPC) codes can provide good power efficiency, which outperform Reed Solomon (RS) codes, convolutional codes and even Turbo codes. LDPC codes usually are decoded by belief propagation, an iterative decoding algorithm. This algorithm typically provides performance near that of the maximum-likelihood decoder, at a much lower complexity. This paper proposes a novel family of quasi-cyclic LDPC codes, which can be applied in deep space missions. Several inequalities are used to delete the short cycles in Tanner graph of LDPC codes, which can affect decoding performance. It extends the class of LDPC codes that can be constructed from shifted identity matrices. The encoders and the decoders are practically feasible because of the quasi-cyclic structure and the inherent block configuration of parity-check matrices. They are linear-time encodable and decodable. The novel LDPC codes perform very well with iterative decoding and demonstrate performance that is better than average random LDPC codes. However, they outperform comparable RS codes etc. This family of quasi-cyclic LDPC codes has three advantages: good performance, low computation complexity and low encoding-decoding complexity. They can be selected for inclusion in a new Consultative Committee for Space Data Systems (CCSDS) telemetry channel coding standard for deep space applications.
- Abstract document
- Manuscript document
IAC-06-B3.2.10.pdf (🔒 authorized access only).
To get the manuscript, please contact IAF Secretariat.