Phil Whiting

Providing quality of service over a shared wireless link

We propose an efficient way to support quality of service of multiple real-time data users sharing a wireless channel. We show how scheduling algorithms exploiting asynchronous variations of channel quality can be used to maximize the channel capacity (i.e., maximize the number of users that can be supported with the desired QoS).

SCHEDULING IN A QUEUING SYSTEM WITH ASYNCHRONOUSLY VARYING SERVICE RATES

We consider the following queuing system which arises as a model of a wireless link shared by multiple users. There is a finite number N of input flows served by a server. The system operates in discrete time t = 0,1,2,…. Each input flow can be described as an irreducible countable Markov chain; waiting customers of each flow are placed in a queue. The sequence of server states m(t), t = 0,1,2,…, is a Markov chain with finite number of states M. When the server is in state m, it can serve mim customers of flow i (in one time slot).The scheduling discipline is a rule that in each time slot chooses the flow to serve based on the server state and the state of the queues. Our main result is that a simple online scheduling discipline, Modified Largest Weighted Delay First, along with its generalizations, is throughput optimal; namely, it ensures that the queues are stable as long as the vector of average arrival rates is within the system maximum stability region.

The Degrees of Freedom of Compute-and-Forward

We analyze the asymptotic behavior of compute-and-forward relay networks in the regime of high signal-to-noise ratios. We consider a section of such a network consisting of K transmitters and K relays. The aim of the relays is to reliably decode an invertible function of the messages sent by the transmitters. An upper bound on the capacity of this system can be obtained by allowing full cooperation among the transmitters and among the relays, transforming the network into a K × K multiple-input multiple-output (MIMO) channel. The number of degrees of freedom of compute-and-forward is hence at most K. In this paper, we analyze the degrees of freedom achieved by the lattice coding implementation of compute-and-forward proposed recently by Nazer and Gastpar. We show that this lattice implementation achieves at most 2/(1+1/K) ≤ 2 degrees of freedom, thus exhibiting a very different asymptotic behavior than the MIMO upper bound. This raises the question if this gap of the lattice implementation to the MIMO upper bound is inherent to compute-and-forward in general. We answer this question in the negative by proposing a novel compute-and-forward implementation achieving K degrees of freedom.

Channel Estimation and Linear Precoding in Multiuser Multiple-Antenna TDD Systems

Traditional approaches in the analysis of downlink systems decouple the precoding and the channel estimation problems. However, in cellular systems with mobile users, these two problems are, in fact, tightly coupled. In this paper, this coupling is explicitly studied by accounting for the channel training overhead and estimation error while determining the overall system throughput. This paper studies the problem of utilizing imperfect channel estimates for efficient linear precoding and user selection. It presents precoding methods that take into account the degree of channel estimation error. Information-theoretic lower and upper bounds are derived to evaluate the performance of these precoding methods. In typical scenarios, these bounds are close.

Scheduling and Pre-Conditioning in Multi-User MIMO TDD Systems

The downlink transmission in multi-user multiple- input multiple-output (MIMO) systems has been extensively studied from both communication-theoretic and information-theoretic perspectives. Most of these papers assume perfect/imperfect channel knowledge. In general, the problem of channel estimation is studied separately. However, in interference-limited communication systems with high mobility, the problem of channel estimation is tightly coupled with the problem of maximizing throughput of the system. In this paper, scheduling and preconditioning in the presence of reciprocal time-division duplex (TDD) training are considered. In the case of homogeneous users, a scheduling scheme is proposed and an improved lower bound on the sum capacity is derived. The problem of choosing training sequence length to maximize net throughput of the system is also studied. In the case of heterogeneous users, a modified pre-conditioning method is proposed and an optimized pre-conditioning matrix is derived. This method is combined with a scheduling scheme to further improve achievable weighted-sum rate.

Computation Alignment: Capacity Approximation Without Noise Accumulation

Consider several source nodes communicating across a wireless network to a destination node with the help of several layers of relay nodes. Recent work by Avestimehr has approximated the capacity of this network up to an additive gap. The communication scheme achieving this capacity approximation is based on compress-and-forward, resulting in noise accumulation as the messages traverse the network. As a consequence, the approximation gap increases linearly with the network depth. This paper develops a computation alignment strategy that can approach the capacity of a class of layered, time-varying wireless relay networks up to an approximation gap that is independent of the network depth. This strategy is based on the compute-and-forward framework, which enables relays to decode deterministic functions of the transmitted messages. Alone, compute-and-forward is insufficient to approach the capacity as it incurs a penalty for approximating the wireless channel with complex-valued coefficients by a channel with integer coefficients. Here, this penalty is circumvented by carefully matching channel realizations across time slots to create integer-valued effective channels that are well suited to compute-and-forward. Unlike prior constant gap results, the approximation gap obtained in this paper also depends closely on the fading statistics, which are assumed to be i.i.d. Rayleigh.

An algorithm for fast, model-free tracking indoors

Tracking people and objects indoors from signal strength measurements has applications as diverse as security monitoring, self-guided museum tours, and personalization of communications services. Accurate dynamic tracking in real-time has been elusive, though, because signal propagation in buildings and the paths that people follow are complex. This paper proposes a simple algorithm for indoor tracking that requires neither a propagation model nor a motion model. It is also simple to compute, requiring only standard tools: Delaunay triangulation, linear interpolation, moving averaging, and local regression. Experiments with real data and simulations based on real data show that the algorithm is not only simple, it is effective.

Optical switch dimensioning and the classical occupancy problem

Results for optical switch dimensioning are obtained by analysing an urn occupancy problem in which a random number of balls is used. This analysis is applied to a high speed bufferless optical switch which uses tuneable wavelength converters to resolve contention between packets at the output fibres. Under symmetric packet routing the urn problem reduces to the classical occupancy problem. Since the problem is large scale and the loss probabilities are small, exact analysis by combinatorial methods is problematic. As an alternative, we outline a large deviations approximation which may be generalised in various ways. Copyright

Optimal resource allocation in HetNets

The deployment of pico cells to cover traffic hot spots within the footprint of a macro cell provides a powerful approach to meet the massive growth in traffic demands fueled by smartphones and bandwidth-hungry applications. Joint optimization of resource allocation and user association is of critical importance to achieve the maximum capacity benefits in such heterogeneous network deployments (HetNets). We first examine the problem of minimizing the amount of resources required to satisfy given traffic demands. We characterize the structure of the optimal solution, and identify a simple optimality condition in terms of the physical transmission rates of the edge users between the macro cell and the various pico cells. We further demonstrate how these structural properties can be leveraged in designing a distributed online algorithm for achieving a max-min fair throughput allocation across all users. Numerical experiments are presented to illustrate the results.

The degrees of freedom of compute-and-forward

We analyze the asymptotic behavior of compute-and-forward relay networks in the regime of high signal-to-noise ratios. We consider a section of such a network consisting of K transmitters and K relays. The aim of the relays is to reliably decode an invertible function of the messages sent by the transmitters. An upper bound on the capacity of this system can be obtained by allowing full cooperation among the transmitters and among the relays, transforming the network into a K × K multiple-input multiple-output (MIMO) channel. The number of degrees of freedom of compute-and-forward is hence at most K. In this paper, we analyze the degrees of freedom achieved by the lattice coding implementation of compute-and-forward proposed recently by Nazer and Gastpar. We show that this lattice implementation achieves at most 2=(1+1=K) ≤ 2 degrees of freedom, thus exhibiting a very different asymptotic behavior than the MIMO upper bound. This raises the question if this gap of the lattice implementation to the MIMO upper bound is inherent to compute-and-forward in general. We answer this question to the negative by proposing a novel compute-and-forward implementation achieving K degrees of freedom.