Coordination of autonomous devices over noisy channels : capacity results and coding techniques

Abstract : 5G networks will be characterized by machine to machine communication and the Internet of Things, a unified network of connected objects. In this context, communicating devices are autonomous decision-makers that cooperate, coordinate their actions, and reconfigure dynamically according to changes in the environment.To do this, it is essential to develop effective techniques for coordinating the actions of the nodes in the network.Information theory allows us to study the long-term behavior of the devices through the analysis of the joint probability distribution of their actions. In particular, we are interested in strong coordination, which requires the joint distribution of sequences of actions to converge to an i.i.d. target distribution in L^1 distance.We consider a two-node network comprised of an information source and a noisy channel, and we require the coordination of the signals at the input and at the output of the channel with the source and the reconstruction. We assume that the encoder and decoder share a common source of randomness and we introduce a state capturing theeffect of the environment.The first objective of this work is to characterize the strong coordination region, i.e. the set of achievable joint behaviors and the required minimal rates of common randomness. We prove inner and outer bounds for this region. Then, we characterize the exact coordination region in three particular cases: when the channel is perfect, when the decoder is lossless and when the random variables of the channel are separated from the random variables of the source.The study of the latter case allows us to show that the joint source-channel separation principle does not hold for strong coordination. Moreover, we prove that strong coordination offers “free” security guarantees at the physical layer.The second objective of this work is to develop practical codes for coordination: by exploiting the technique of source polarization, we design an explicit coding scheme for coordination, providing a constructive alternative to random coding proofs.
Document type :
Theses
Complete list of metadatas

Cited literature [71 references]  Display  Hide  Download

https://tel.archives-ouvertes.fr/tel-02269626
Contributor : Abes Star <>
Submitted on : Friday, August 23, 2019 - 10:11:24 AM
Last modification on : Wednesday, September 11, 2019 - 2:19:38 PM

File

69871_CERVIA_2018_diffusion.pd...
Version validated by the jury (STAR)

Identifiers

  • HAL Id : tel-02269626, version 1

Collections

Citation

Giulia Cervia. Coordination of autonomous devices over noisy channels : capacity results and coding techniques. Other. Université de Cergy Pontoise, 2018. English. ⟨NNT : 2018CERG0960⟩. ⟨tel-02269626⟩

Share

Metrics

Record views

168

Files downloads

26