Research efforts of the UCLA Communication Systems Laboratory are focused on developing practical coding schemes that approach information theoretic limits. Our work applies information theory, communication theory, and signal processing to analyze and design algorithms to improve the throughput, robustness, and reliability of communications systems.
One current area of interest is variable-length coding schemes that employ feedback, which can allow rates to closely approach capacity at short blocklengths [ISIT'12], [ITW'12]. Furthermore, the use of rate-compatible codes in an incremental redundancy setting can often allow simpler decoding operations as compared to a single long-blocklength code.
The lab has a long history of low-density parity check (LDPC) code design work, including LDPC coding for flash memories [Globecom'11], LDPC code construction via an absorbing set spectrum approach [ISIT'11], design of rate-compatible protograph-based raptor-like (PBRL) LDPC codes [Globecom'11], and design and analysis of nonbinary LDPC codes [Globecom'11].
Past network coding projects include universal recovery [Allerton'11] and file transfer in peer-to-peer networks [Allerton'11]. Other recent projects have included nonlinear trellis coding for multiple-access channels [TCOM'12], superposition coding in the degraded broadcast channel [MILCOM'11], optimal allocation of packet-level erasure coding and physical-layer channel coding [TCOM'11], and multiterminal source coding [ISIT'12].