2008 /   2009  /   2010 /   2011
1. Special Lectures 1           2. Special Lectures 2

2009

This series of lectures aims to develop an understanding of acoustics and audio in relation to the human listener. Five basic topics are introduced, each with a focus on how they relate to humans: acoustics, signal processing, psychoacoustics, spatial hearing and audio systems. In acoustics, we focus on how sound behaves in the human environment, including the effects of room acoustics and noise. In audio signal processing, we overview basic tools of digital audio. In psychoacoustics, we introduce the theory of loudness and other psychoacoustical parameters. In spatial hearing, we examine the mechanisms for sound localization and other spatial percepts. In audio systems, we examine basic audio systems, and how such systems can be evaluated. The final two lectures examine how the principles from the earlier lectures can be applied in sonification design - which is the representation of data to people using non-speech audio. The lectures may be particularly useful for students who have an interest in human application of sound, acoustics and audio systems.

Research in communication networking is about efficient sharing of network resources such as link bandwidth and router buffer space. The research problem can be formulated as an optimization problem to achieve optimal performance where the performance metric can be delay, throughput, fairness, etc. The complexity of solving the optimization depends on whether the problem is linear or non-linear, constrained or unconstrained, convex or non-convex, etc. In this lecture series, linear programming (LP) will first be introduced since many network optimization models actually are special types of linear programming problems. A remarkably efficient solution technique, the simplex method, will be described. The concept of duality and its many important ramifications in linear programming will also be discussed since the relationships between the dual problem and the original primal problem prove to be extremely useful in many situations. The second part of the lecture will introduce non-linear programming (NLP) since some network optimization problem is non-linear. The optimality conditions, known as the Karush-Kuhn-Tucker (KKT) conditions will be discussed and the concept of duality will also be introduced. Applications of LP and NLP with specific examples in network optimization will be illustrated. The first application is introduced through the network flow problem which is formulated as a linear programming problem. The third part of the lecture series will continue to discuss about applications specific to network optimization. The example will be drawn from the bandwidth sharing problem. The concept of fairness and Pareto efficiency will be introduced. The optimal bandwidth sharing problem will be formulated as a Network Utility Maximization (NUM) problem. Through the KKT conditions and the duality theory, a distributed method to obtain the optimal solution will be presented to illustrate the important application of duality theory. Students will have the opportunity throughout the course to use the optimization solver, Lingo, to solve some of problems discussed in the course.