- About EECS
- The EECS Experience
- Research Overview
- Research Areas
- Research Facilities
- Technology Transfer
- Research Opportunities for Undergraduates
- Online Library
- Industry Relations
Once every week while school is in session, EECS invites a distinguished researcher or practitioner in a computer science or electrical and computer engineering-related field to present their ideas and/or work. Talks are generally targeted to Electrical Engineering and Computer Science graduate students. This colloquium series is free and open to everyone.
Monday, March 10, 2014 - 8:45am - 10:00am
Software surrounds us and drives our lives: most modern systems heavily rely on software. Managing the complexity of these large software systems is a challenging task. It is even more challenging for software product lines -- families of software variants with similar yet not identical functionality (think Samsung's Galaxy S line which includes over two dozen smartphone models, each with its own variation of software).
In the last two decades, numerous approaches have been proposed to help develop product lines in an efficient manner. Surprisingly, they are rarely used in practice. In this talk, we investigate reasons for such lack of adoption. We then propose an approach for improving existing practices that relies on a set of software analysis and transformation tasks.
We show that some of the required tasks are barely studied by existing works and look in detail into one of them: matching artifacts from multiple products (a.k.a. n-way match). We explain why the n-way match problem is NP-hard, explore possible approximate solutions, and propose our own heuristic n-way match algorithm that is superior to other practical approaches in terms of its accuracy. We conclude the talk by outlining a future research agenda.
Tuesday, March 11, 2014 - 8:45am - 9:45am
Location-based services such as targeted advertisement, geo-social networking and emergency services, are becoming increasingly popular for mobile applications. While on-board sensors on the smartphones such as GPS are able to provide accurate outdoor locations, accurate indoor localization schemes now still require either additional infrastructure support (e.g., ranging devices) or extensive training before system deployment (e.g., WIFI signal fingerprinting). In this talk, I will talk about our latest work that uses wireless access point sequence as a new metric for fingerprinting-based indoor localization systems. This metric is resilient to time varying WIFI signal changes, heterogeneous devices and dynamic power control of wireless access points, while is able to achieve very good system performance. Based on this metric, we designed the first signal-fingerprinting indoor localization system that is able to automatically construct the finger-print map and completely eliminate the heavy training.
Another topic I will cover in this talk is large-scale battery management. Large-scale batteries have been widely adopted in applications such as electric vehicles and energy storage in power grids. While the improvement of the battery energy density is relatively slow in the past decade, in this talk, I will discuss how the exploration of battery cell reconfigurations at large-scale battery systems can benefit from real-time dynamic controls for both discharging and charging processes. The experimental results with commercial battery cells on our customized testbed, as well as EV-trace driven emulations demonstrate significantly improved energy efficiency of our proposed designs.
Finally, I'll briefly talk about the convergence of mobile device energy management and battery energy management, and how battery-aware energy management can help prolong the mobile device lifetimes in realistic environments.
Wednesday, March 12, 2014 - 8:45am - 9:45am
Safety-critical embedded systems, e.g., avionics, automotive, and medical devices, must tightly integrate and coordinate embedded computing systems with physical elements in a timely and dependable fashion. The current design process leverages results from the real-time scheduling theory, which considers tasks or jobs (from the operating system concept of thread) as the units for the analysis and validation. As a result, timing is often considered as a “non-functional” requirement which will only be checked after the system integration, while it should be a correctness criterion starting from the functional design. In addition, the constantly growing complexity of embedded systems coupled with the tight cost and short time-to-market often results in long design iterations to improve the design and fix errors, and ultimately sub-optimal solutions.
We propose to make time a first-class citizen of system design, and consider timing in the design synthesis from the functional models. Different from the traditional research in real-time systems community, the task (or threads) model becomes an intermediate artifact, and the timing analysis becomes part of a synthesis problem. We will focus on the Synchronous Reactive (SR) model, since it is very popular for modeling safety-critical embedded applications. We will automate the design optimization and synthesis of automotive systems that go from system-level modeling to correct, predictable, and efficient implementation. The implementation will be targeted at all kinds of practical architecture platforms, including single-core, multi-core, time-triggered distributed systems, and distributed systems without synchronized clocks.
Monday, March 17, 2014 - 8:45am - 10:00am
Coding is like gardening; it requires good plan, good supplies, but most importantly continuous nurture and maintenance.
In this talk, I will concentrate on refactorings and program transformations that help nurture good code by removing code smells and vulnerabilities. I will describe OpenRefactory/C, an infrastructure for building program transformations for C programs. C, in spite of its popularity, has IDEs with a limited portfolio of program transformations, with limited scalability and limited applicability to real-world programs. OpenRefactory/C aims to have full support for the C preprocessor, support for static analyses, and an API and environment that make it easy for new developers to contribute new refactorings. Refactorings that we have implemented on OpenRefactory/C are bug-free, unlike the refactorings featured in commercial IDEs such as Eclipse CDT, Visual Studio, etc.
I will also describe three complex, security-oriented program transformations that fix issues in C integers. These transformations fixed all variants of integer vulnerabilities featured in benchmark programs of NIST's SAMATE reference dataset and 5 open source software, making the changes automatically on over 15 million lines of code. Being integrated with source code and development process, refactorings and program transformations not only help maintain good code, but also teach developers about how to write and appreciate good code.
Tuesday, March 18, 2014 - 8:45am - 9:45am
A defining feature of a smart grid is its ability to incorporate advanced data analytics for real time monitoring and control. However, heavy reliance on networking for data collection inevitably exposes a grid to threats of cyber attacks. In this talk, we consider man-in-the-middle attacks on power system topology and state estimation, where an attacker alters certain meter data to mislead the control center with an incorrect network topology or state estimate. First, we present a necessary and sufficient condition under which an undetectable attack exists. When an undetectable attack is feasible, we develop a data-driven attack mechanism that does not require any system parameter information; we show that partial meter observations are sufficient for designing undetectable attacks. Then, in order to protect a grid from potential attacks, we develop countermeasures based on meter data authentication. It is shown that if data from a set of meters satisfying a certain graph-covering property are protected, any attack can be detected.
Wednesday, March 19, 2014 - 8:45am - 10:00am
Advancements in silicon technology, computer architecture, wireless communications, and computer networking have led to the proliferation of embedded systems in a plethora of application domains (e.g., industrial and home automation, automotive, space, medical, defense, etc.). To meet the diverse requirements of these application domains as well as to permit sophisticated applications of greater value as compared to an isolated embedded system, embedded systems are often networked together and processing is done in parallel within an embedded system to form parallel and distributed embedded systems. A crucial enabler for these parallel and distributed embedded systems is burgeoning multi-core revolution in computing industry.
This seminar discusses architecture, modeling, and optimization of parallel and distributed embedded systems with an emphasis on multi-core. There has been an increasing proliferation of diverse multi-core architectures to keep up the Moore's law, which necessitates evaluation of these architectures to determine the most befitting architecture for an application. The seminar elaborates the evaluation of multi-core architectures with two of the speaker's research contributions in this domain. The first contribution proposes a queueing theoretic approach for modeling multi-core architectures that provides a quick and inexpensive performance evaluation both in terms of time and resources. The second contribution evaluates two embedded multi-core architectural paradigms: symmetric multiprocessors (SMPs) and tiled multi-core architectures (TMAs), based on parallelized benchmarks.
The seminar then discusses three application domains for parallel and distributed embedded systems: embedded wireless sensor networks, cyber-transportation systems, and IP multimedia subsystem outlining the speaker's key contributions in each application domain. The contributions aim at the architecture, modeling and optimization for embedded systems' design metrics, viz., performance, power, dependability, and security for these application domains. In particular, the seminar elaborates a multi-core-based approach for the design of secure and dependable cybercars (next generation of automobiles) with steer-by-wire as a case study. The seminar culminates with the speaker's future research agenda in parallel and distributed embedded computing domain.
Thursday, March 20, 2014 - 8:45am - 9:45am
Vast and deep integration of renewable energy resources into the existing power grid is essential in achieving the envisioned sustainable energy future. Many countries around the globe as well as many states in the U.S. have set up aggressive Renewable Portfolio Standards (RPSs). The state of California, as an example, has targeted a 33% RPS by 2020. Volatility, stochasticity, and intermittency characteristics of renewable energies, however, present a challenge for integrating these resources into the existing grid in a large scale as the proper functioning of an electric grid requires an instantaneous and continuous power balance between supply and demand. In this talk, we see how the demand-side flexibility can be used to match random supply. In particular, we show that the thermal storage potential of a collection of Thermostatically Controlled Loads (TCLs) (such as residential air conditioners) can be modeled as a battery and can be leveraged for enabling a deep penetration of renewable energy resources.
Wednesday, April 2, 2014 - 8:45am - 10:00am
Mobile devices are a viable platform to help people live a safer, smarter and more productive life. Not only do they enable ubiquitous access to online services, but also offer services on the go, such as healthcare monitoring and public sensing. However, the increasing resource demands of mobile services and the inherent constraints of mobile devices limit the quality and functionality that can be offered, preventing mobile devices from exploiting their full potential as reliable service providers. In this talk, I will introduce two concepts that will revolutionize mobile services: cloud-assisted service provisioning and service personalization. Cloud-assisted service provisioning bridges the gap between limited resources of mobile devices and increasing resource demands of mobile services and applications. Service personalization capitalizes on context information and user preferences to personalize mobile services and provide superior user experience. The talk will present the status quo in this domain and outline future research plans.
Friday, March 7, 2014 - 8:45am - 10:00am
Department of Computer Science
University of Texas at Austin
Thursday, March 6, 2014 - 8:45am - 10:00am
Robert Bosch Research and Technology Center North America
Wednesday, March 5, 2014 - 8:45am - 9:45am
Harpreet S. Dhillon
Postdoctoral Research Associate
Department of Electrical Engineering
University of Southern California
Monday, March 3, 2014 - 8:45am - 9:45am
Research Assistant Professor
Department of Electrical and Computer Engineering
University of Minnesota
Friday, February 28, 2014 - 8:45am - 10:00am
Computer Science and Engineering Department
University of Texas at Arlington
Monday, February 24, 2014 - 4:00pm - 4:50pm
University of Minho / Oregon State University
Thursday, February 20, 2014 - 8:45am - 9:45am
Directions to Oregon State and parking information
The colloquium series takes place on the main Oregon State campus. The colloquium is held usually in KEC 1001, 4:00 - 4:50 pm. [map and catalog listing]. However, please be sure to check the schedule, as room changes occur with some frequency.
Speaking in the colloquium series
Speaking in the EECS Colloquium series provides an excellent opportunity to meet and interact with faculty and students of the Oregon State School of EECS. Our faculty are always interested in exploring new possibilities of collaboration. If you would be interested in being invited to speak, please contact the colloquium director.
Visiting the School of EECS before or after a colloquium
If you are planning on attending a colloquium, consider taking the opportunity to visit the School of EECS before or after the colloquium. To make arrangements for your visit, please contact Tina Batten via e-mail or phone (541-737-8613).
Groups attending the colloquium series
If you intend to bring a substantial size group (10 people or more), please notify the colloquium director so that we can ensure that we schedule a lecture room with sufficient capacity.
For more information
School of EECS
Oregon State University
1148 Kelley Engineering Center
Corvallis, OR 97331-5501
Phone: +1 541-737-4544