 |
Tutorials 6 June 2005
Tutorial 1
Optimal Subspace-Based Signal Processing Outline In this tutorial a new tool called the subspace perturbation expansion (SPE) is introduced. The SPE is a formula that shows, to first-order, how much the subspaces (e.g. signal subspace and its orthogonal complement) of a finite-size data matrix are perturbed by additive noise. This formula is more general that an asymptotic Taylor series expansion of eigenvectors of a covariance matrix. Because the formula is first-order in the noise, it is valid when the signal-to-noise ratio (SNR) is high enough. The SPE may be used to evaluate the performance of existing subspace algorithms, and to derive new algorithms. The new algorithms are statistically efficient (estimator variance reaches the CR bound) over the range of SNRs in which the SPE is valid. This SNR range is the same as that required by all other subspace algorithms. Several new algorithms will be presented. All of them are derived from the SPE formula and they produce estimates the variances of which achieve the CR bound for the problem. The first problem that will be discussed is sensor-array processing for direction-of-arrival (DOA) estimation. For shift-invariant structures we derive a new algorithm called Optimal Subspace Estimation (OSE) and for problems without shift invariance (e.g. for arbitrary arrays) we use the SPE to derive a weighted subspace fitting (WSF) algorithm for estimating the DOAs. The second problem that will be discussed is that of estimating the parameters of signals consisting of exponentially damped sinusoids in noise. The final topic of the tutorial is matrix filtering in which we develop the design equations for matrix filters and show how to incorporate matrix filters into OSE and WSF.
Tutorial 2
Iterative ('Turbo') Algorithms on Factor Graphs Outline This tutorial describes turbo algorithms in the framework of factor graphs.
Half Day Mon 6 June 13.00-16.15
Ultrawideband Standards UWB Course:
Ultrawideband technologies are about to become popular throughout the world, and will usher in data rates never before conceived in wireless networks. UWB will enable real time video and USB Cable replacement, and is already being designed into products by major consumer equipment (CE) manufacturers. This course presents information about the evolution and implementation of the Ultrawideband (UWB) standard, IEEE 802.15.3.a, and provides real-world viewpoints regarding the uses, applications, and economic constraints for UWB in emerging commercial applications. The course also presents the widely varying viewpoints held by international regulatory bodies regarding the use of shared spectrum, and highlights recent rulings by European spectrum bodies that may hamper worldwide adoption of UWB.
|
|||
| |||
