Packet loss and delay call for adaptive and error-robust coding schemes for multimedia traffic. Many proposed scalable coding formats require complex bit stream manipulation at intermediate locations in the network. Such requirements are unrealistic in present network nodes, so new solutions are needed. Emerging network handling techniques such as prioritization, intelligent packet dropping and congestion control feedback can be integrated into coding schemes in order to change this situation. We focus our work on scalable and robust source coding techniques that exhibit graceful degradation under error-prone conditions. In general, these coding schemes belong to the adaptive source coding domain which includes layered coding, scalable coding and multiple descriptions (MD). These will be investigated in combination with packet prioritization, rate adaptivity and differentiated packet dropping policies.

Project coordinator: Professor Andrew Perkis