| Under the background of multimedia communications over multiple-path networks, we propose several communications model and schemes to simulate the practical multimedia communications. We utilize the rate-distortion theory and classic Gaussian model to formulate our problem. The end-to-end average distortion is our performance evaluation criterion. Also we use the optimization techniques (convex optimization, linear programming) to solve our optimization problems, that is, to attain the minimum average distortion. Specifically, we focus on Multiple Description Coding (MDC) for correlated multi-path networks.;For erasure channels, we compare MDC and SDC and solve the rate allocation problem. And for fading channels, we work out the distortion exponent for general number of channels and asymmetric SNR, and linear programming formulation is proposed for the distortion exponent of MDC for general number of descriptions. Also we analytically characterize the effect of correlation and side information on the distortion exponent in terms of power offset. Finally the achievable distortion exponent for MDC over interference channel is investigated. |
