Grant: Discovery Grants Program - Individual

In the past decade, relay-assisted communications have been the center focus of intensive research studies. Numerous relying strategies, including amplify-and-forward, decode-and-forward, estimate-and-forward, have been proposed in the context of one-way or two-way relay-networks with the aim to enable communications between one or more pairs of source-destination or to enable information exchange between one or more pairs of transceivers. It turns out that the relay channel belongs to a larger class of channels, namely active channels, where the channel behavior can be adjusted, for optimal performance, somewhere between the two ends of the channel. Such channel adjustability does not exist in traditional channel models (which they are herein referred to as passive channels). Indeed in traditional models of wireless channels, the end-to-end channel is often given and there is no control on the channel behavior. To compensate the channel, one has to either resort to precoding at the transmitter and/or equalize the channel at the receiver. In active channels, the channel itself can be adjusted in addition to signal precoding at the transmitter and/or channel equalization at the receiver. For example, in a relay channel, the end-to-end channel impulse response can be carefully determined through judiciously designing the relaying operation and optimizing a certain performance metric over the channel characteristics and over the signal precoding operation and receiver side channel equalization. The applicant has pioneered the research in active channels through several publications which appeared in prestigious periodicals, namely IEEE Transactions on Signal Processing (two papers) and IEEE Signal Processing Letters (one paper). In these published results, the problem of sum-rate maximization has been studied for i) one-way active parallel channels with equal sub-channel noise powers, ii) one-way active parallel channels with unequal subchannel noise powers, and iii) two-way active parallel channels with equal noise powers.In all these research outcomes, there is one common assumption: the received signal-to-noise ratios (SNRs) over different sub-channels are linearly related to the channel gains squared. Based on this assumption, then the sum-rate is maximized over the transmit powers of the source in one-way schemes (or the transmitter powers of the two transceivers over different sub-channels in two-way schemes). In this proposal, the applicant details a research program that allows optimal resource allocation for active channels. This is achieved by introducing more general (nonlinear) models to describe the sub-channel SNRs and sub-channel gains. Such a task requires thisprogram to focus on studying different performance measures. This research program will study parallel channels with the aim to increase the data rate beyond current network data rates.