Sibi Raj B. Pillai

On the sum capacity of multiaccess block-fading channels with individual side information

We consider the problem of finding optimal, fair and distributed power-rate strategies to achieve the sum capacity of the Gaussian multiple-access block-fading channel. The transmitters have access to only their own fading coefficients, while the receiver has access to all of the fading coefficients. We propose a distributed strategy called the ‘midpoint’ strategy which is optimal when the system cannot tolerate outage. In addition, we demonstrate a successive decoding scheme that can achieve this maximal sum-rate. In presence of outage, we show that the strategies based on a single threshold are suboptimal.

Extensions to Orthogonal Matching Pursuit for Compressed Sensing

Compressed Sensing (CS) provides a set of mathematical results showing that sparse signals can be exactly reconstructed from a relatively small number of random linear measurements. A particularly appealing greedy-approach to signal reconstruction from CS measurements is the so called Orthogonal Matching Pursuit (OMP). We propose two modifications to the basic OMP algorithm, which can be handy in different situations.

On the noisy feedback capacity of Gaussian broadcast channels

It is well known that, in general, feedback may enlarge the capacity region of Gaussian broadcast channels. This has been demonstrated even when the feedback is noisy (or partial-but-perfect) and only from one of the receivers. The only case known where feedback has been shown not to enlarge the capacity region is when the channel is physically degraded. In this paper, we show that for a class of two-user Gaussian broadcast channels (not necessarily physically degraded), passively feeding back the stronger users signal over a link corrupted by Gaussian noise does not enlarge the capacity region if the variance of feedback noise is above a certain threshold.

On the adaptive sum-capacity of fading MACs with distributed CSI and non-identical links

We consider a two-user block-fading MAC with distributed channel state information (CSI), where each user has access to only its own fading coefficients. The average rate-pairs of communication while employing within-block coding is known as the adaptive capacity region, where each user adapts the rate based on its perceived link gain. We evaluate the adaptive sum-capacity of MAC channels with general fading distributions, for discrete as well as continuous valued ones.

Distributed Rate Adaptation and Power Control in Fading Multiple Access Channels

Traditionally, the capacity region of a coherent fading multiple access channel (MAC) is analyzed in two popular contexts. In the first, a centralized system with full channel state information at the transmitters (CSITs) is assumed, and the transmit power and data-rate can be jointly chosen for every fading vector realization. On the other hand, in fast-fading links with distributed CSIT, the lack of full CSI is compensated by performing ergodic averaging over sufficiently many channel realizations. Notice that the distributed CSI may necessitate decentralized power-control for optimal data-transfer. Apart from these two models, the case of slow-fading links and distributed CSIT, though relevant to many systems, has received much less attention. In this paper, a block-fading additive white Gaussian noise MAC with full CSI at the receiver and distributed CSI at the transmitters is considered. The links undergo independent fading, but otherwise have arbitrary fading distributions. The channel statistics and respective long-term average transmit powers are known to all parties. We first consider the case where each encoder has knowledge only of its own link quality, and not of others. For this model, we compute the adaptive capacity region, i.e., the collection of average rate-tuples under blockwise coding/decoding such that the rate-tuple for every fading realization is inside the instantaneous MAC capacity region. The key step in our solution is an optimal rate allocation function for any given set of distributed power control laws at the transmitters. This also allows us to structurally characterize the optimal power control for a wide class of fading models. Further extensions are also proposed for the case where each encoder has additional partial CSI about the other links.

Distributed Scheduling Schemes in Energy Harvesting Multiple Access

We consider a time-slotted distributed multiple access channel model with energy harvesting transmitters, where each transmitter is equipped with a battery of finite capacity. At the start of each time slot, a random amount of harvested energy is available from an exogenous energy source. Each transmitter only knows its own arrival process, and is unaware of the energy arrival processes at the other terminals, leading to a distributed access system. For this set up, we design efficient rate and power scheduling policies to maximize the average throughput (sum-rate) under slot-wise coding and decoding requirements. We show that the proposed policies achieve the same gap to optimality as the recently proposed ergodic averaging schemes which require coding and decoding over a large number of slots in our setting.

Ordered Orthogonal Matching Pursuit

Compressed Sensing deals with recovering sparse signals from a relatively small number of linear measurements. Several algorithms exists for data recovery from the compressed measurements, particularly appealing among these is the greedy approach known as Orthogonal Matching Pursuit (OMP). In this paper, we propose a modified OMP based algorithm called Ordered Orthogonal Matching Pursuit (Ordered OMP). Ordered OMP is conceptually simpler and provides an improved performance when compared to OMP.

On the Ergodic Sum-Capacity of Decentralized Multiple Access Channels

We consider a fast fading AWGN multiple-access channel (MAC) with full receiver CSI and distributed CSI at the transmitters. The objective is to evaluate the ergodic sum-capacity of this decentralized model, under identical average powers and channel statistics across users. While an optimal water-filling solution can be found for centralized MACs with full CSI at all terminals, such an explicit solution is not considered feasible in distributed CSI models. Our main contribution is an upper-bound on the ergodic sum-capacity when each transmitter is aware only of its own fading coefficients. Interestingly, our techniques also suggest an appropriate lower bound. These bounds are shown to be very close to each other, suggesting the tight nature of the results.

On fading MAC channels with asymmetric CSI

We consider a distributed MAC setting with block-wise flat fading links and full receiver CSI (channel state information). Of the L transmitters, a subset is assumed to have knowledge of the global CSI vector in each block, whereas the remaining users have access only to their respective link qualities, i.e. each one in the latter set is unaware of the quality of other links. Outage is not allowed in any communication block. We propose efficient power-allocation and rate-adaptation strategies which are sum-rate optimal when users in each subset observe identical fading distributions chosen from a class, which includes the popular Rayleigh, Ricean etc.

Decentralized power adaptation in ergodic fading multiple access channels

We consider decentralized power control in fading multiple-access channels (MAC). When compared to a centralized MAC, solving the optimal power control for a decentralized system is difficult even for moderate number of users. Good thumb-rules for decentralized power control are available in literature, which are also asymptotically tight in the number of users. In this paper, we first derive a structural property that any optimal decentralized power control scheme must follow, and use this property to suggest a modified power control scheme. Numerical results suggest that the scheme we propose outperforms the existing ones.