Minimizing The Expected Distortion Using Multi-layer Coding Approach In Two-Hop Networks (1605.05988v2)

Abstract: This work concerns minimizing the achievable distortion of a Gaussian source over a two-hop block fading channel under mean square-error criterion. It is assumed that there is not a direct link between transmission ends and the communication is carried out through the use of a relay, employing the Decode and Forward (DF) strategy. Moreover, for both of transmitter-to-relay and relay-to-destination links the channel statistics are merely available at the corresponding receivers, while the channel gains are not available at the affiliated transmitters. It is assumed that a Gaussian source is hierarchically encoded through the use of successive refinement source coding approach and sent to the relay using a multi-layer channel code. Similarly, the relay transmits the retrieved information to the destination through the use of a multi-layer code. Accordingly, in a Rayleigh block fading environment, the optimal power allocation across code layers is derived. Numerical results demonstrate that the achievable distortion of the proposed multi-layer approach outperforms that of single-layer code. Moreover, the resulting distortion of DF strategy is better than Amplify and Forward~(AF) strategy for channel mismatch factors greater than one, while the resulting distortion has a marginal degradation for channel mismatch factors lower than one.