Optimal linear-combining receiver for decode-and-forward relays using superposition coding

We consider a relay network with a single source, a single destination, and multiple relays. The relays are half-duplex and use the decode-and-forward protocol. A layered Gaussian source partitioned into two layers is transmitted. The source broadcasts a message consisting of two layers to all the relays using superposition coding. Each relay re-encodes and forwards all its successfully decoded layers to the destination using an orthogonal channel. Superposition coding is used only if both layers are forwarded. We propose a novel approach to derive the optimal linear-combining receiver at the destination. Numerical results show that the two-layer scheme using superposition coding outperforms a conventional one-layer scheme in terms of average throughput and expected distortion.