MIMO two-way X relay channel, and L-cluster MIMO multiway relay channel. Previous studies mainly considered the achievability of the DoF cut-set bound 2N at the antenna configuration N < 2M by applying signal alignment for network coding. This work aims to investigate the achievability of the Do

Abstract not found

Abstract—Duty cycling mode is widely adopted in wireless sensor networks to save energy. Existing duty-cycling protocols cannot well adapt to different data rates and dynamics, resulting in a high energy consumption in real networks. Improving those protocols may require global information or heavy

mcmxciv This thesis deals with combinatorics in connection with Coxeter groups, finitely generated but not necessarily finite. The representation theory of groups as nonsingular matrices over a field is of immense theoretical importance, but also basic for computational group theory, where the group elements are data structures in a computer. Matrices are unnecessarily large structures, and part of this thesis is concerned with small and efficient representations of a large class of Coxeter groups (including most Coxeter groups that anyone ever payed any attention to.) The main contents of the thesis can be summarized as follows. • We prove that for all Coxeter graphs constructed from an n-path of unlabelled edges by adding a new labelled edge and a new vertex (sometimes two new edges and vertices), there is a permutational representation of the corresponding group. Group elements correspond to integer n-sequences and the nodes in the path generate all n! permutations. The extra node has a more complicated action, adding a certain quantity to some of the numbers.

Abstract not found

ACKNOWLEDGMENTS I would like to give many thanks to Dr. Shengli Fu and Dr. Yan Huang as my advi-

de nationalité suisse et originaire de Zurich (ZH) et Lucerne (LU) acceptée sur proposition du jury:

algorithm constrained split augmented Lagrangian shrinkage algorithm for balanced model (C-SALSA-B) converges faster than previously proposed algorithms accelerated proximal algorithm (APG) and alter-nating directional method of multipliers for balanced model (ADMM-B).

Abstract — We consider minimum variance estimation within the sparse linear Gaussian model (SLGM). A sparse vector is to be estimated from a linearly transformed version embedded in Gaussian noise. Our analysis is based on the theory of reproducing kernel Hilbert spaces (RKHS). After a characterization of the RKHS associated with the SLGM, we derive a lower bound on the minimum variance achievable by estimators with a prescribed bias function, including the important special case of unbiased estimation. This bound is obtained via an orthogonal projection of the prescribed mean function onto a subspace of the RKHS associated with the SLGM. It provides an approximation to the minimum achievable variance (Barankin bound) that is tighter than any known bound. Our bound holds for an arbitrary system matrix, including the overdetermined and underdetermined cases. We specialize

We explore two fundamental questions at the intersection of sampling theory and information theory: how is channel capacity affected by sampling below the channel’s Nyquist rate, and what sub-Nyquist sampling strategy should be employed to maximize capacity. In particular, we first derive the capacity of sampled analog channels for two prevalent sampling mechanisms: filtering followed by sampling and sampling following filter banks. Connections between sampling and MIMO Gaussian channels are illuminated based on this analysis. Optimal prefilters that maximize capacity are identified for both cases, as well as several kinds of channels for which these sampling mechanisms are optimal to maximize capacity at sub-Nyquist rates. We also highlight connections between sampled analog channel capacity and minimum mean squared error estimation from sampled data. In particular, it is shown that for both filtering and filter-bank sampling strategies, the filters maximizing capacity and minimizing mean squared error are equivalent. We also investigate a more general sampling strategy by adding modulation banks to filter-bank sampling. This general sampling method subsumes most nonuniform sampling techniques applied in both theory and practice. We also show a connection between this general sampling method and MIMO Gaussian channels. We then identify the optimal sampling strategy for piece-wise flat sampled