- Principal investigator: Suhas Diggavi (UCLA)
- CO-PI: Ruediger Urbanke
- Sponsor: Swiss national science foundation
- Supported students: Sanket Dusad, Etienne Perron
- Project timeline: 2006-2008
In this project we examined fundamental questions about opportunistic coding and their applications. We examined the question in two contexts. One on channel coding techniques suitable for wireless communication. The other on source coding coding ideas applicable to wireless sensor networks.
Rate and reliability form a fundamental trade-off in communications. A higher rate comes at the cost of lower reliability and vice versa. By utilizing the randomness in the channel realizations, we designed opportunistic coding strategies that utilize good channels without a-priori knowledge about them. We established fundamental bounds on such coding schemes and also made progress in the design of practical opportunistic codes. The particular focus of this project was on broadband wireless communication, where there was significant inter-symbol interference. This project was an extension of the earlier project on narrow-band diversity embedded codes, whose description can be found here
This was part of a larger effort on opportunistic communication and diversity embedded codes, details of which can be found here.
We also extended the philosophy of opportunistic coding to multi-terminal source coding. We investigated the fundamental limits of lossy source coding which opportunistically utilizes the presence of multiple processes (side-information) correlated with the source. Some of this work can be found here.