Joint Source-Channel Coding

While the transmission of information to be recovered with the best end-to-end fidelity is a joint source-channel coding problem in general, Shannon's well-known separation theorem \cite{Shannon:1948} reduces it to separate compression and channel encoding problems without losing the end-to-end optimality when this transmission is between a point to point link (under certain technical conditions. See  LINK to SEPARATION THEOREM). Inspired by this result, the design of transmission protocols for wired systems has traditionally followed the division of the transmission between the nodes in the network into non interfering point to point links and a block architecture that provides modularity.

However, the optimality of separation does not necessarily generalize to all systems. While from an information theoretic perspective, optimal end-to-end performance can be achieved by separate source and channel coding in point to point static channels; this optimality breaks down


 * 1) when the channel is fading and the application has delay limitations;
 * 2) in multi-user networks;
 * 3) when the complexity of the system is limited, that is, in almost all practical systems.

See , and CITE, CITE for some examples proving the suboptimality of separate source and channel coding in multiuser systems, and ,, for examples of fading scenarios in which source-channel separation is not optimal.

While assumptions 1.,2., and 3., are reasonable assumptions for wired links, and separation approach has allowed to design systems with high performances in wired systems, this approach has become a bottleneck in the design of wireless transmission schemes due to the broadcast and the highly varying nature of the wireless channel.


 * Basic transmission schemes
 * Uncoded transmission
 * Separate Source-channel coding
 * Multi-layer Transmission (Broadcast approach)
 * Hybrid Digital-analo Transmission
 * Delay Limited systems
 * Time-varying Channels
 * Delay-free schemes


 * Transmission of Correlated sources over Multiuser scenarios
 * Broadcast Channel (BC)
 * Multiple Access Channel (MAC)
 * Relay Channel
 * Interference Channel (IFC)