Link to
Colorado State University Home Page

BMAC Spring 2007: Abstracts


Kinematically redundant robots - The promise of human-like dexterity
Tony Maciejewski
Department of Electrical and Computer Engineering
Colorado State University

The vast majority of robots in use today operate in very structured environments, e.g., in factory assembly lines, and possess only those limited motion capabilities required to perform specific tasks. While these robots can outperform humans in terms of speed, strength, and accuracy for these tasks, they are no match for the dexterity of human motion. Part of a human's inherent advantage over industrial robots is due to the large number of degrees of freedom in the human body. Articulated, i.e., jointed, motion systems that possess more degrees of freedom than the minimum required to perform a specified task are referred to as kinematically redundant. In an effort to mimic the dexterity of biological systems, researchers have built a number of kinematically redundant robotic systems, e.g., anthropomorphic arms, multi-fingered hands, dual-arm manipulators, and walking machines. While these systems vary in their appearance and intended applications, they all require motion control strategies that coordinate large numbers of joints to achieve the high degree of dexterity possible with redundant systems. This talk will discuss the issues that arise when designing such strategies, frequently drawing on the use of the singular value decomposition, including the characterization of redundancy, the quantification of dexterity, and the development of efficient and numerically stable motion control algorithms that simultaneously optimize multiple criteria. In addition, the ability of kinematically redundant robots to sustain component failures and yet still complete an assigned task will addressed, thereby extending the application of robots to environments that are unacceptable or inaccessible to humans.

BIO

Tony Maciejewski received the PhD degree in Electrical Engineering from The Ohio State University and served on the faculty at Purdue University for 14 years. He is currently the Head of Electrical and Computer Engineering at Colorado State University. He is a Fellow of the IEEE for his contributions to the design and control of kinematically redundant robots.


Split Decomposition of Graphs
Ross McConnell
Department of Computer Science
Colorado State University

The split decomposition of a graph is a well-known, unique recursive decomposition of a graph. It has many beautiful combinatorial properties that were worked out in the early 1980's.

The decomposition has applications to solving combinatorial problems that have important industrial applications, such as finding a maximum independent set of vertices in a graph. The strategy works as follows: the decomposition identifies compartmentalized regions of the graph whose interactions with the rest of the graph have a simple structure. When these regions are small, as they always are on some well-known classes of graphs, an inefficient algorithm can be applied independently to each of these regions. Due to the simple relationships between the regions, these results can then be combined by induction to get an efficient solution for the whole graph.

The talk will start with a tutorial on the elegant and intuitive properties of the decomposition, and finish with some related algorithms that my students Ben Joeris, Scott Lundberg, and I have recently discovered. The talk is aimed at a broad audience, and should be accessible to anyone who is comfortable with induction and recursion.

BIOs

Ben Joeris is in his second year as an undergrad at CSU, and this past November, he was one of four Americans to rank in the top 48 in the world in the annual TopCoder International Algorithms Competition. He has been working with McConnell since he was a junior at Fort Collins High, and is currently spending a semester abroad studying algorithms and combinatorics in Budapest. Scott Lundberg got his B.S. in computer science at CSU, and is currently completing his master's degree while working at Numerica, Inc. McConnell earned his Ph.D. at the University of Colorado in Boulder in 1994, and has been an assistant professor at CSU since 2002. He has taught at the University of Colorado at Denver, Willamette University and Amherst College, and has worked at Vexcel Corporation in Boulder developing algorithms and scientific software for the Jet Propulsion Lab.


The Economic Performance of Dot-Com Firms and the Impact of the Internet on Traditional Business: Two Case Studies from the History of the Commercialized Internet
William Aspray
Professor of Informatics
Indiana University

This paper will present two case studies from a forthcoming book, William Aspray and Paul Ceruzzi, eds., The Internet and American Business (MIT Press, late 2007). One case study will discuss the market strategy and structure of dot-com start-ups and compare their performance with that of companies in manufacturing and various high-tech industries. The other case study will examine the impact of the Internet on a traditional business sector. This business sector has the largest revenue stream of any legal segment on the Internet (but you will have to come to the talk to find out what that business sector is!). The talk will also discuss how one frames the story of the development of the Internet since its commercialization in 1991.

BIO

William Aspray is Rudy Professor of Informatics at Indiana University in Bloomington. He has formerly held teaching positions at Harvard, Minnesota, Penn, Rutgers, and Williams; and he has worked in various administrative positions, including as executive director of Computing Research Association. His research involves the history, policy, and social study of information technology. His best-known books are John von Neumann and the Origins of Modern Computing (MIT Press, 1990) and Computer (Basic Books, 1996; 2nd edition 2004; with Martin Campbell-Kelly). His most recent books are Globalization and the Offshoring of Software (ACM, 2006, edited with Frank Mayadas and Moshe Vardi) and Women and Information Technology (MIT Press, 2006, ed. With Joanne Cohoon). Current research includes work on IT entrepreneurship, the impact of digitalization on the media industries, and the informatics of diabetes.


Cost-sensitive Classifier Evaluation Using Cost Curves
Robert C. Holte
Department of Computing Science
University of Alberta
(joint work with Chris Drummond
National Research Council, Ottawa)

The evaluation of classifier performance in a cost-sensitive setting is straightforward if the operating conditions (misclassification costs and class distributions) are fixed and known. When this is not the case, evaluation requires a method of visualizing classifier performance across the full range of possible operating conditions. This talk argues that the classic technique for classifier performance visualization -- the ROC curve -- is inadequate for the needs of researchers and practitioners in several important respects. It then describes a different way of visualizing classifier performance -- the cost curve -- that overcomes these deficiencies. No familiarity with ROC curves or cost curves is necessary, they will be fully explained.


Vulnerabilities and Opportunities in SMS-Capable Cellular Networks
Patrick McDaniel
Department of Computer Science and Engineering
The Pennsylvania State University

Cellular networks are a critical component of the economic and social infrastructures in which we live. In addition to voice services, these networks deliver alphanumeric text messages to the vast majority of wireless subscribers. To encourage the expansion of this new service, telecommunications companies offer connections between their networks and the Internet. The ramifications of such connections, however, have not been fully recognized. In this talk, we evaluate the security impact of the SMS interface on the availability of the cellular phone network. Specifically, we demonstrate the ability to deny voice service to cities the size of Washington D.C. and Manhattan with little more than a cable modem. Moreover, attacks targeting the entire United States are feasible with resources available to medium-sized zombie networks. This analysis begins with an exploration of the structure of cellular networks. We then characterize network behavior and explore a number of reconnaissance techniques aimed at effectively targeting attacks on these systems. We conclude by discussing counter-measures that mitigate or eliminate the threats introduced by these attacks, and identify opportunities and requirements for the security infrastructure of the next generation cellular networks.

This work was reported in New York Times, Reuters, The Associated Press, and many other venues of the popular press. For more detail: http://www.smsanalysis.org/

BIO

Patrick McDaniel is the Hartz Family Career Development Assistant Professor in the Computer Science and Engineering Department at the Pennsylvania State University, and co-director of the Systems and Internet Infrastructure Security Laboratory. He received his Ph.D. from the University of Michigan in 2001 where he studied the form, algorithmic limits, and enforcement of security policy. Prior to joining Penn State, Patrick was a senior technical staff Member of the Secure Systems Group at AT&T Labs-Research and Adjunct Professor of the Stern School of Business at New York University.

Patrick's recent research efforts have focused on telecommunications security, distributed systems security, network security, language- based security, and public policy and technical issues in digital media. Patrick is a past recipient of the NASA Kennedy Space Center fellowship, a frequent contributor to the IETF security standards, and has authored many papers and book chapters in various areas of systems security. He is the co-chair of the 2007 and 2008 IEEE Symposium on Security and Privacy, and served as the Program Chair of the 2005 USENIX Security Symposium, the Vice Chair for Security and Privacy for WWW 2005, and is the Chair of the Industry and Government Track at the 2005 and 2007 ACM Computer and Communications Security conference. Patrick is also an associate editor of the journals IEEE Transactions on Software Engineering and ACM Transactions on Internet Technologies . Prior to pursuing his Ph.D. in 1996, Patrick was a software architect and program manager in the telecommunications industry.


Turning Software Development into Income: An Example
David Schwaab
Rebit Co.
Fort Collins, Colorado

Mr. Schwaab will present on the keys and pitfalls of creating business from software development, using his current experience with the Rebit Company as an example. The presentation will focus on the pragmatic aspects of creating software businesses, with an emphasis on addressing any questions and aspects of software business development in which the audience has interest.

BIO

David Schwaab has lived on the Front Range of Colorado since 1969, and has divided his career between Denver and the Silicon Valley. He is a principal of two successful software startups, and worked in marketing and business management at Hewlett-Packard for 26 years. He is currently a principal of the Rebit Company, a high-tech startup based in Loveland, CO. Mr. Schwaab attended CSU and Regis College, and holds a BS in business administration.


Malicious Transactions in Mobile Database Systems
Prof. Vijay Kumar
Computer Science & Informatics
University of Missouri-Kansas City

Securing the database from the effects of malicious activities and maintaining data consistency have never been easy. This task becomes significantly more complex in Mobile Database Systems (MDS) because of the unique demands it imposes on data processing activity. The characteristics of a malicious transaction have been discussed in many papers; however, a formal definition seems to be missing. Such definition is very useful, rather essential to investigate its interaction with MDS and to develop schemes for its management. In this seminar, we ponder over and discuss a few things about malicious transactions and its interactions with MDS. In particular, we first try to develop a formal definition of malicious transaction, explain our reference architecture of a mobile database system, define the structure and processing of mobile transactions, and investigate interaction of malicious transactions with mobile database systems. Finally, we present an outline of a scheme, which we call "Location Signature", to identify the attack of malicious transactions on mobile database systems.

BIO

Dr. Vijay Kumar is a professor of Computer Science at the University of Missouri-Kansas City. His research activities, at present, are bounded by (a) Data Dissemination through Wireless Channels, (b) Mobile Database Systems (MDS), (c) Sensor Technology, (d) System Security and Trust management, (e) Web Technology, and (f) Bioinformatics. A number of these projects are funded by grants from National Science Foundation, HP labs, Sprint, and St Lukes Hospital research Foundations. He has 70+ refereed publications in ACM TODS, IEEE TKDE, IEEE TMC, ACM Communications, and others. He has written four books which have been published by John Wiley, Prentice Hall, and Kluwer.


MANET Simulation Studies: Reversing the Incredibles
Tracy Camp
Department of Computer Science
Colorado School of Mines

A mobile ad hoc network (MANET) is defined by mobile network nodes, no fixed infrastructure, multi-hop communication, and unreliable wireless links. In the MANET research community, simulation is the main tool used to study these dynamic networks. To determine the state of MANET simulation studies, we surveyed the 2000-2005 proceedings of the ACM International Symposium on Mobile Ad Hoc Networking and Computing (MobiHoc). From our survey, we found significant shortfalls in the credibility of presented simulation results.

A simulation scenario must be properly constructed to be effective in evaluating the performance of a MANET routing protocol. For example, a poorly defined routing protocol might appear successful in a scenario with a low average hop count because little routing is needed. Of course, many standards are needed to establish a rigorous evaluation procedure for MANET simulation research, e.g., standards are needed for simulation scenarios, random number generation, and results analysis.

In this talk, I will propose and explore two standard metrics that researchers should consider to ensure long routes are available and used in the evaluation of a MANET routing protocol. I will also discuss tools that my students and I have developed to assist MANET researchers in improving the credibility of MANET simulation-based studies. Lastly, I will briefly present past successes, current research challenges, and future directions of my research group.

BIO

Tracy Camp is an Associate Professor of Computer Science at the Colorado School of Mines. She is the Founder and Director of the Toilers (http://toilers.mines.edu), an active ad hoc networks research group currently consisting of five faculty and 18 students. Dr. Camp has received 12 grants from NSF, including a CAREER award in 1997. She has published over 50 refereed articles and 9 invited articles; as of August 2006, her articles have been cited over 1300 times (per Google Scholar). Dr. Camp recently returned from New Zealand, where she was a Fulbright Scholar. She is currently a member of both the Ad Hoc Networks Journal and the Pervasive and Mobile Computing Journal editorial boards, the elected Treasurer of ACM's Special Interest Group on Mobile Computing (SIGMOBILE), and a member of ACM's Committee on Women in Computing. Dr. Camp was recently invited to be an ACM Distinguished Lecturer (August 2006), awarded IEEE Senior Member status (July 2006), and selected as an ACM Distinguished Scientist (October 2006).


Careers and Challenges for Computer Scientists at Sandia National Laboratories
Jean-Paul Watson
Senior Member of Technical Staff
Sandia National Laboratories

The US Department of Energy's national laboratories provides an alternative career path for Computer Scientists, who generally focus on industrial and academic positions. This talk will survey the history and role evolution of the US national laboratory system, with specific emphasis on Sandia National Laboratories, a large (10,000 employee) engineering laboratory headquartered in Albuquerque, New Mexico. I will discuss the unique career challenges faced by a relatively new PhD (myself) in this environment, as contrasted with academic and industrial positions. I will also survey a number of key problems I have tackled - and helped to solve - over my relatively short tenure at Sandia. Interactivity, especially by soon-to-be graduating students, will be encouraged.

BIO

Dr. Jean-Paul Watson graduated from Colorado State University with a PhD in Computer Science in 2003, and was advised by Professors Adele Howe and Darrell Whitley. He is currently a Senior Member of Technical Staff at Sandia National Laboratories, with responsibilities in research and management of applications ranging from water security, military logistics, materials science, and fundamental optimization algorithms.


Using Random Circulations to Find Small Cuts
Ramakrishna Thurimella
Department of Computer Science
University of Denver
(joint work with David Pritchard, University of Waterloo)

Using basic properties of circulations, we show that it is easy to generate random circulations on G which can be used to compute the cut edges, cut edge-pairs, and cut vertices of a graph. For example, the cut edges are "usually" the edges where a random circulation vanishes. In the distributed setting, this leads to an improvement in the time complexity of finding cut edge-pairs to O(Diam), where Diam is the diameter of the graph . Asymptotic improvements are also implied for finding cut vertices. These algorithms are the Las Vegas kind and use messages of length O(log |V |). The distributed cut vertex algorithm can also be used to find the blocks of G.

BIO

Ramki Thurimella received his Ph.D. from the University of Texas at Austin in 1989. He spent 2 years as a Research Associate at the Institute for Advanced Computer Studies at the University of Maryland, before joining the University of Denver in 1991, where he is an Associate Professor in the Department of Computer Science. His research interests include algorithms, networks security, intrusion detection, reputation systems, Internet and peer-to-peer computing. He is the recipient Research Initiation and other grants from NSF.


Genuinely Secure Systems
Bill Worley
CTO
Secure64 Software Corporation

Secure64 Software Corporation has developed a micro-OS called "SourceT." SourceT was designed from the ground up to provide the foundation for systems with a strong set of security properties. Such systems are called "Genuinely Secure." The Intel Itanium architecture for the first time provides capabilities that enable a genuinely secure system. The talk will define and discuss the properties of a genuinely secure system, the innovative uses of unique Itanium processor capabilities, and the synergy with other emerging trusted system directions. Properties of Secure64's first product based upon the SourceT micro-OS also will be presented.

BIO

Dr. William S. Worley Jr. (Bill)
CTO, Secure64 Software Corporation
Retired HP Fellow, Chief Scientist, and Distinguished Contributor
Former Commissioner, Colorado Governor's Science and Technology Commission Education

MS physics, University of Chicago
MS Information Science, University of Chicago
PhD Computer Science, Cornell University

Bill is a system architect. At Hewlett Packard he directed the team that developed the PA-RISC architecture. He later directed the development of the PA Wide Word architecture - the foundation for the HP/Intel partnership that led to the Itanium 2 microprocessor family. Prior to HP, during 13 years with IBM, he had contributed to architectures for mainframes, storage systems, and IBM's first RISC architecture. In the years prior to his retirement from HP Bill focused upon hardware and software architectures for secure systems. Following retirement, Bill joined Secure64 Software as a co-founder and CTO. Secure64 has developed a multi-core platform control system, including a queued, asynchronous network stack, which fully exploits the security capabilities of the Itanium architecture. This foundation is inherently secure by design, self-defending, high performance, and scalable. It will feature growing capabilities for integrated defensive countermeasures and applications executing under securely supported guest operating systems.


Why Robots Will Never Have Sex
Avinash Kak
Electrical and Computer Engineering Department
Purdue University

Despite all the ambiguities associated with our understanding of "intelligence", many futurists would have you believe that the day is fast approaching when the machines will be far more intelligent than humans. Some of wilder projections made by such individuals include scenarios in which robots take over the world and enslave the humans. The goal of my talk will be to point to the self-limiting nature of rationality and thus also the self-limiting nature of any intelligence frameworks based on rationality. Rational frameworks for reasoning, including those that take uncertainty into account, are closed systems of mutually-consistent formulas that are antithetical to what it means to be a human. So if the machines of the future will never be human-like in their thinking abilities, what will they be? Is there an end-point in the evolution of robotic intelligence (that unlike human intelligence is easily quantifiable). To answer this question, I will focus on sensory intelligence for the machines of the future. In particular, I will present examples of the progress that has been in robotic vision. While this progress has certainly made robotic vision useful for many industrial and societal tasks, in and of itself the robotic vision systems are merely encodings of the solutions produced by the human mind for solving specific and extremely narrowly defined problems.

That brings me to the title of this talk. While obviously a cheap hook, it is nevertheless intended to convey the possibility that it is our emotions, our passions, our innate desires --- all ingredients of our sexuality -- that are the defining elements of our consciousness and, through consciousness, our intelligence.

BIO

Avinash Kak is a Professor of Electrical and Computer Engineering at Purdue University. His research and teaching include sensor networks, computer vision, robotics, and high-level computer languages. At this time, he is spearheading a large research effort is camera-based sensor networks for habitats of the future. His latest book "Programming with Objects", published by John Wiley and Sons in 2003, is used by a number of leading universities as a text on object-oriented programming. His forthcoming book "Scripting with Objects" focuses on object-oriented scripting. These are two of the three books for an "Objects Trilogy" he is creating. The last, expected to be finished sometime in 2008, will be titled "Designing with Objects". His coauthored book "Principles of Computerized Tomographic Imaging" was republished as a Classic in Applied Mathematics by SIAM (Society of Industrial and Applied Mathematics). His other co-authored book "Digital Picture Processing" is also considered by many to be a classic in computer vision and image processing.


A Platform-Independent Component QoS Modeling Language for Distributed Real-Time and Embedded Systems
Aniruddha Gokhale
Electrical Engineering and Computer Science Department
Vanderbilt University

Distributed real-time and embedded (DRE) systems require multiple, simultaneous quality of service (QoS) properties, such as predictability, reliability and security. When DRE systems are assembled from components, it becomes necessary for system designers to ensure that the system composition and its QoS configurations are functionally and systemically compatible. Unfortunately, different QoS properties of DRE systems tend to conflict with each other, and thereby require design time QoS tradeoff analysis. This talk describes a model-driven engineering approach to analyze the validity of the system's QoS configurations. First, we describe the design of a Component QoS Modeling Language (CQML), which enables a clean, visual separation of the multiple QoS aspects while internally maintaining a unified view of the system. Second, we illustrate how well-defined formalisms of system behavior defined in CQML are leveraged to conduct a real-time schedulability analysis in the presence of multiple QoS requirements.

BIO

Aniruddha Gokhale is an Assistant Professor of Computer Science and Computer Engineering in the Dept. of Electrical Engineering and Computer Science at Vanderbilt University. He joined the Institute for Software Integrated Systems at Vanderbilt University in Jan 2002 as a Research Scientist and became a faculty member in Sept 2003. He was previously with Lucent Bell Labs in Murray Hill, NJ. His research at Vanderbilt focuses on applying model driven engineering techniques for the deployment, configuration, optimization and QoS provisioning of middleware systems. At Vanderbilt his research has led to the open source model driven middleware tool suite called CoSMIC. His research has been funded by DARPA, industry and NSF.

Dr. Gokhale received his Bachelors (BE) from University of Pune, India in 1989; the Masters (MS) in Computer Science from Arizona State University in Summer 1992; and PhD in Computer Science from Washington University in St. Louis in May 1998.