Distributed gain matrix optimization in non-regenerative MIMO relay networks

Abstract

We consider a half-duplex multiple-input multipleoutput relay network with multiple source and relay nodes and a set of collocated destination antennas. Each relay is equipped with multiple antennas and is constrained to map its receive signal linearly to its transmit signal. We devise a gradient based scheme that enables each relay to locally optimize its gain matrix with respect to the achievable sum-rate without requiring explicit knowledge of all fading coefficients characterizing the channel state. In particular, we demonstrate by numerical evidence that the overhead due to exchange of channel state information required by the scheme does either not scale at all or, if so, very slowly with the number of relay nodes or antennas therein.

Universal computation with low-complexity wireless relay networks

“…We obtain a solution by applying a modified version of the gradient based algorithm presented in [10], where decentralized relay gain allocation has been employed in a communications context. We can accordingly calculate the complex gradient of the norm in (10) Here, 0* denotes the complex conjugate, OR is the Henni tian transposition, and Ek is a matrix with all zeros except for a one on the kth diagonal position.…”

MIMO relaying with compact antenna arrays: Coupling, noise correlation and superdirectivity

“…The optimization problem can be stated formally as: For solving the optimization problem in (19), we use a gradient-search algorithm similar to the one in [15]. Due to the different power constraint, the Jacobian matrix in [15] is modified to incorporate (15).…”

“…7, the antenna separation distances at the relay and at the destination are 0.1λ. For reference, we plotted the rates given that the relay uses spatially white transmission, i.e.Ĝ = cI NS , where c is selected to satisfy the relay power constraint in (15).…”

“…Among different schemes for relay gain allocation, the gradient based ones are of particular relevance to our work [15], [16], [17]. While [15], [16] assume full channel state information knowledge, [17] uses the knowledge of statistical properties of the channels in the optimization.…”

“…While [15], [16] assume full channel state information knowledge, [17] uses the knowledge of statistical properties of the channels in the optimization. These studies do not consider the stated physical effects.…”