Technical Program


In the early days of ‘discrete time signal processing’ in the 1950s it was of mostly theoretical interest owing the cost and complexity of the electronics needed to implement it. It was first used in defense applications, such as radar signal processing, where the performance improvements it offered offset that cost and complexity, and in sonar and audio applications where the signal bandwidth was low enough to match the low speed at which digitizers and computers could run. The world is wildly different now. We note regularly that signal processing ‘is everywhere’ and that ‘DSP inside’ stickers should be applied to virtually every device that you own, including your car and refrigerator. This radical shift over the past 50 years is the focus of the technical program for this joint DSP and Signal Processing Education workshop. The early theoretical fascination with the field through roughly the year 2000, coupled with rare and/or expensive implementation, has seemingly been replaced with ubiquitous application of DSP, enabled by the incredible improvement in semiconductors, coupled with declining interest in DSP as a research area. We examine this paradoxical situation via the technical program for this workshop by viewing it from three perspectives ---

By providing all three of these perspectives, the intent of the technical program is to provoke examination of the field and its future by the participants of the workshop. The fact that signal processing is going to be ‘inside’ of virtually everything means that signal processing education must remain vibrant and improve, and that SP-based solutions need to be developed for a wide range of the world’s problems and for an even wider range of commercial opportunities. How to do that best is in our hands.

Plenary Presentations

Large-scale video processing for YouTube

Dr. Michele Covell, Senior Staff Research Scientist at Google Research

DSP System Design Challenges

Dr. James Truchard, President, CEO, and Cofounder of National Instruments

Intelligent hyperspectral imaging and remote sensing from space using onboard digital signal processing

Dr. Jeff Puschell, Principal Engineering Fellow at Raytheon Space and Airborne Systems

Computers that understand speech: Where are we? Where are we going?

Dr. Roberto Pieraccini, CEO International Computer Science Institute

Invited Talks, Panels, and Special Sessions

Panel: Signal Processing at the University: Still an Active Research Area or ‘Just’ an Undergraduate Service Course?
Professor Ronald Schafer, Stanford University
Professor Lawrence Rabiner, Rutgers University
Professor Gene Franz, Rice University
Gene, Larry, and Ron all have both academic and industry experience. Armed with these dual perspectives and given their long and successful careers from the dawn of DSP to now , they will discuss the future of signal processing, both analog and digital, in the university education process, ranging from a focus on ground-breaking research to the need for broad SP education to support other technical fields and the engineers entering industry. Their initial presentations will be followed by modulated (moderated?) discussion, debate, and questions from the workshop attendees.

Invited Talk: Predictable Impacts of the Development of Effective Signal Processing MOOCS
C.S. Burrus, Dean emeritus, Rice University
The rapid rise of Massive Open On-Line Courses has evoked considerable interest and controversy regarding their applicability to signal processing education. Sidney Burrus will report on recent efforts to bring laboratory experience to a DSP MOOC and then will extend the discussion to extrapolate the long term impacts of this educational approach. MOOCs will be presented as one element in the expanding array of Open Educational Resources (OER). General discussion and moderated debate will follow his presentation.

Invited talk: Digital pre-distortion of power amplifiers in cellular wireless systems
Dr. Chris Dick, DSP Chief Architect, Communications Signal Processing Group, Xilinx

In the last 4 years there has been explosive growth in mobile IP usage. ITU ICT statistics report that in the period from 2000 to 2010 the number of mobile cellular subscriptions increased by 8 fold, a rate that astonishingly exceeds the growth in the number of internet users over the same period. In response to the insatiable demand for wireless capacity, network operators are rapidly deploying wideband waveforms such as multi-carrier LTE-A, for example 4-carrier LTE signals occupying 100MHz of transmission bandwidth. But capacity is not the only dimension of the problem, cost and power efficiency are key parameters. With the RF-shelf comprising a large fraction of the bill-of-materials, the task of the designer naturally falls to reducing the cost of the RF processing sub-system. With the power amplifier at the center of the tradeoff between efficiency, linearity, transmission bandwidth and transmission power, digital pre-distortion (DPD) has become a core technology in every basestation, and is very active area of research of commercial wireless radio labs.

This presentation begins with a review of the motivation behind the extreme focus that every manufacturer of cellular wireless equipment, from small-cell to macro-cell, has on DSP-based signal conditioning of LTE-A, W-CDMA, CDMA2000, GSM and multi-protocol transmission waveforms. We then provide an overview of DSP-based DPD and describe how DPD is realized in a basestation. Linearization results are presented for multi-carrier LTE-A signals based on a real-time DPD implementation on a Xilinx Zynq SoC device using a basestation class Doherty power amplifier.

Invited Talk: Large data, stream processing & project-based learning
Dr. Darel Linebarger, Sr. Manager, SPC System Design, The Mathworks, Inc.

In this talk we provide an overview of the latest trends for signal processing in MATLAB which includes new features to handle large data, to facilitate processing of real-time streaming signals, and to enhance project-based learning by targeting low-cost hardware. We will showcase MATLAB test benches that make it easier to connect to hardware and stream real-time data for processing into MATLAB. For example, the Microsoft Kinect can be used as a recording device for audio and video testbenches. We will also highlight new features in MATLAB for authoring dynamic systems and algorithms that operate on streaming data, including system objects that provide the algorithmic components as well as tools to visualize execution of dynamic systems.

Invited Talk: Signal Processing Initiatives from the SPS Education Committee
Professor Douglas B Williams, Georgia Institute of Technology
Progress and future directions for Connexions, Sig Port, Sig View (aka SPS Tutorial Program), and SP Wikipedia.

As part of the IEEE Signal Processing Society's Membership Board, the Education Committee is associated with several projects aimed at providing educational materials for the Society's members as well as the general public. This talk will overview those projects and provide information on how to become involved. Current projects include the IEEE-CNX Connexions Project, SigPort, SigView (aka the SPS Tutorial Project), and the SP WikiProject.

Special attention will be paid to the Society's WikiProject. Wikipedia has become one of the most widely accessed resources on signal processing, used frequently by researchers, teachers, and students. Maintaining a high quality standard for signal processing articles on Wikipedia is important in defining how our field is viewed by the public. However, the freelance-editing model of Wikipedia makes quality control and maintenance of these articles a real challenge. As researchers in the field, there is much that we can do towards improving and maintaining the quality of such articles without too much effort. Moreover, editing an article in Wikipedia can be expected to have markedly more impact and visibility than writing a paper on the topic; for example, the page on Nyquist–Shannon sampling theorem is viewed on average a 1000 times a day.

The Education Committee's action plan will be presented for building a Wikiproject on Signal Processing comprising a group of volunteers working together to improve SP articles on Wikipedia. There will then be opportunities later in the workshop to participate in sessions for evaluating and modifying signal processing content in Wikipedia.

Invited Talk: Transitions
Professor Gene Franz, Rice University

Reflections on the transition, after 39 years in industry, to Professor in the Practice at Rice University in the context of the birth, development, and maturation of DSP.

Tuesday Afternoon Workshops

Distributed Special Interactive Sessions: Signal Processing Wikipedia
Small group participation in evaluation of current content, assessment of needs, generation of new content in:

Please contact Professor Doug Williams (doug.williams@ece.gatech.edu) with questions about the workshop.

Understanding Learning Difficulties in Continuous-time Signals and Systems Courses and Making These Courses More Accessible
Mario Simoni and Maurice Aburdene

This workshop will be an engaging opportunity to address two questions:
  1. Why do students struggle so much with concepts in continuous-time signals and systems (CTSS) courses?
  2. What are some methods that instructors can use to help the students grasp the fundamental concepts?
The introductory CTSS course is one of the most difficult courses that students encounter in an electrical and computer engineering (ECE) curriculum, as evidenced by well-above-average drop/failure rates. To help explore the problem, we will spark discussion by presenting data from surveys, focus groups, historical data, the CTSS concept inventory, and the Index of Learning Styles. To demonstrate effective methods for engaging students, hands-on activities that were developed at Rose-Hulman and Bucknell will be presented.

Please contact Professor Mario Simoni (simoni@rose-hulman.edu) with questions about the workshop.

Technical Program Schedule

Program Schedule
Poster Sessions