Poster Session and Reception

Poster Sessions 2011

There will be a poster session and reception from 3:40 p.m. to 5:00 p.m. on January 27th, in CSL 301.

________________________________________________________________________________________

Organization: University of Illinois at Urbana-Champaign
Poster Session and Reception in CSL 301

Abstract:

A P300 speller is a Brain-computer interface that can be used as a means of communication for people with impaired sensory motor function. This speller is described by different parameters such as the number of illuminations and the inter-stimulus interval. The performance of P300 spellers is not easily quantifiable, and therefore there is often no systematic way of choosing the system parameters. To correct this problem, we propose a Markov chain model of the P300 speller. This model allows us to quantify performance, and subsequently choose the system parameters in an optimal fashion.

Biography:

Navid Aghasadeghi is currently pursuing his Ph.D. degree in Electrical Engineering under the supervision of Dr. Timothy Bretl. He obtained his BS degree in 2008 from the University of Texas at Austin, and his MS degree in 2010 from the University of Illinois at Urbana Champaign. His research interests lie in control theory, neuroscience and brain-machine interfaces.

_______________________________________________________________________________________________________________

Organization: University of Illinois at Urbana-Champaign
Poster Session and Reception in CSL 301

Abstract:

The classical problem of robot coverage is to plan a path that brings a given point on the robot within some fixed radius of every point in the workspace. We are interested in a variant of this problem in which we are only required to cover a portion of the workspace along its boundary but must deal with significant uncertainty in both sensing and actuation. In particular, we consider a differential-drive robot with access only to wheel encoders and to boundary sensors that trigger when entering or leaving the workspace. We capture uncertainty by assuming drift in the wheel sizes and wheel base of this robot. We show that it is possible to estimate these parameters based on the difference in time between predicted and observed boundary crossings, but only for a good choice of intermediate trajectory. We apply a linear analysis of observability to design a sequence of these trajectories that allow both calibration and coverage near the boundary. Finally, we validate our approach using experiments both in simulation and with a real robot.

Biography:

Colin Das is fascinated by robotic coverage. His recent work focuses on the design of motion plans to guarantee local coverage under uncertainty. He is also extending these ideas to developing simultaneous calibration and coverage techniques with minimal sensing. He received his B.S. in Aerospace Engineering from the University of Illinois at Urbana-Champaign with a focus in control systems.

_______________________________________________________________________________________________________________

Organization: University of Illinois at Urbana-Champaign
Poster Session and Reception in CSL 301

Abstract:

We consider a dynamic zero-sum game between two players. The first player acts as a controller for a discrete time LTI plant, while the second player acts to jam the communication between the controller and the plant. The number of jamming actions is limited. We determine saddle-point equilibrium control and jamming strategies for this game under the full state, total recall information structure for both players, and show that the jammer acts according to a threshold policy at each decision step. Various properties of the threshold functions are derived and complemented by numerical simulation studies.

Biography:

Abhishek Gupta is currently pursuing his PhD in Aerospace Engineering in University of Illinois at Urbana-Champaign. He has completed his Bachelor of Technology from Indian Institute of Technology, Bombay, India in 2009. His current research interests are in Distributed controls and security issues in Networked Control Systems.

_______________________________________________________________________________________________________________

Organization: University of Illinois at Urbana-Champaign
Poster Session and Reception in CSL 301

Abstract:

The optimal steady state operation of the vapor compression cycle (VCC) is analyzed from the perspective of the underlying physics of the cycle. Three degrees of freedom of the VCC are optimized using two different objective functions---one that is based on the first law of thermodynamics and the other which minimizes the rate of exergy destruction in the cycle. The use of exergy is motivated by its ability to capture the physics of both the first and second law of thermodynamics in a single property. The optimization variables are chosen as the three enthalpies, h1, h2, and h3=h4, of the cycle, as well as the refrigerant mass flow rate. A truck transport refrigeration system (TTRS) is considered as a case study on which the optimization is applied. The exergy-based objective function produces similar results as the energy-based objective function. The optimal set points generated by the exergy-based objective function are also shown to provide an increase of 33% in COP compared against the nominal set points regularly used for the actual TTRS being considered. The optimization results highlight the evaporator and condenser pressures and temperatures as critical parameters in improving the efficiency of the overall cycle operation.

Biography:

Neera Jain received her B.S. from the Massachusetts Institute of Technology in 2006 and her M.S. from the University of Illinois at Urbana-Champaign in 2009, both in mechanical engineering. She also completed teaching certification for secondary (8-12) mathematics in the state of Massachusetts. She is currently a Ph.D. candidate in mechanical engineering at the University of Illinois with a focus in control systems. She was awarded the Eugene and Lina Abraham Endowed PhD Supplemental Fellowship in 2009. Neera is currently a DOE Office of Science Graduate Research Fellow; her research focuses on the development of algorithms to optimize the performance and reduce energy consumption of different energy and power systems.

_______________________________________________________________________________________________________________

Organization: University of Illinois at Urbana-Champaign
Poster Session and Reception in CSL 301

Abstract:

This work considers the problem of trajectory tracking for a UAV under model uncertainty. A standard approach to trajectory tracking is receding-horizon LQR. However, as the model becomes more inaccurate we expect performance to degrade. One alternative that addresses this is L1 adaptive control. In this work, we consider a different approach based on reinforcement learning. We begin by considering trajectories as finite sequences of motion primitives. One approach to tracking such trajectories is to compute a feedback policy for each trajectory, such as finite- or infinite-horizon LQR, and switch between these policies at the appropriate time. However, because of large model inaccuracies, tracking along a single primitive can exhibit poor performance and even instability. When tracking a sequence of primitives and switching between controllers, these problems can be amplified because of growing state error at each switching point. Recent work in apprenticeship learning demonstrates techniques that improve the tracking of particular trajectories. We use a similar approach to learn improved feedback policies for a finite set of primitives, and study the affect of this learning on switching stability.

Biography:

Miles Johnson is a PhD candidate in Aerospace Engineering at the University of Illinois at Champaign-Urbana. He obtained a B.S. from Cornell University, after which he worked on the imaging team on the Mars Exploration Rover missions. Currently, he works on applying principles from robotics and control theory to design better neural interfaces. In particular, his research uses non-invasive neural sensors such as electroencephalography (EEG) and electromyography (EMG) to provide control over vehicles with complex dynamics, for example wheelchairs, cars, and aircraft.

_______________________________________________________________________________________________________________

Organization: University of Illinois at Urbana-Champaign
Poster Session and Reception in CSL 301

Abstract:

Stretchable sensor arrays have numerous applications in robotics, structural health monitoring (SHM), and electronic textiles. The ability to fabricate thin-film transistor (TFT) active devices for stretchable sensor arrays opens up new possibilities in data acquisition and conditioning. TFT active devices for stretchable sensor arrays require fabrication on polymer substrates, which introduces new process challenges due to thermal limitations, which ultimately limits the TFT performance. In this research, we demonstrate low-temperature fabrication processes for amorphous oxide semiconductor (AOS) TFTs, using room-temperature RF sputtering for deposition of InGaZnO and Silicon Nitride thin-films on a polyimide substrate. These devices have comparable performance to AOS TFTs fabricated on rigid substrates, having an electron mobility exceeding 10cm2/ (V s), and an on-off current ratio of 106.

Biography:

Kevin Lin obtained his B. S. in Electrical Engineering and Computer Science at University of California-Berkeley 2006 and M.S. in Electrical and Computer Engineering at University of Illinois at Urbana-Champaign 2008. Currently, he is a PhD student in the department of Electrical and Computer Engineering, University of Illinois at Urbana-Champaign.

_______________________________________________________________________________________________________________

Organization: University of Illinois at Urbana-Champaign
Poster Session and Reception in CSL 301

Abstract:

Details coming soon

Biography:

Gayathri is pursuing her PhD in the Mechanical Engineering department, working with Prof. Srinivasa Salapaka. She received her Bachelors from Anna University, Chennai in 2006 and her Masters from UIUC in 2009. Her research interests include nonlinear control and application of control techniques to atomic force microscopy.

_______________________________________________________________________________________________________________

Organization: University of Illinois at Urbana-Champaign
Poster Session and Reception in CSL 301

Abstract:

Brain computer interfaces (BCIs) have been creating a direct link between human thought and computer input for nearly 40 years. However, these interfaces are often slow and difficult to use because of the noisy, low-bandwidth, nature of measurements of brain activity. This is particularly true in non-invasive techniques such as electroencephalography (EEG), which measures electrical potentials in the brain. In this work, we will use perceptual feedback to enhance the strength of one such potential, the steady state visually evoked potential (SSVEP). The SSVEP is a natural electrical entrainment in the primary visual cortex as a response to a repetitive visual stimulus modulated at 3.5 to 75Hz. This phenomenon is largely considered to be a passive response, but it has been shown that the strength of the signal can be modified through feedback. In our experiment, we use an audible tone as feedback to the participant, with frequency proportional to the strength of the SSVEP. The participant learns to control their SSVEP through controlling the audible tone. We demonstrate the efficacy of this technique and discuss it's potential for increasing the performance of SSVEP-based BCIs.

Biography:

James Norton is a native Floridian from Daytona Beach. He pursued his undergraduate degree in psychology at the University of Florida. While in Gainesville, he participated in research relating to emotional expression in Parkinson's disease. For the past three years, Jamie has been working as environmental consultant and Web programmer for Abt Associates, an EPA contractor, in Cambridge, Massachusetts. Currently, Jamie Norton is a PhD student in the Neuroscience program and an affiliate of the NSF’s Neuroengineering IGERT program at the University of Illinois.

_______________________________________________________________________________________________________________

Organization: University of Illinois at Urbana-Champaign
Poster Session and Reception in CSL 301

Abstract:

Details coming soon

Biography:

Cem Onyuksel is a first-year graduate student in the Electrical and Computer Engineering department at the University of Illinois at Urbana-Champaign. He received his B.S. in Electrical and Computer Engineering from Carnegie Mellon University in May 2010. His research interests include robotics, control, artificial intelligence, and brain-machine interfacing.

_______________________________________________________________________________________________________________

Organization: University of Illinois at Urbana-Champaign
Poster Session and Reception in CSL 301

Abstract:

Given a deterministic control system, the problem of inverse optimal control is to find a cost functional with respect to which an observed control law is optimal. In solving this problem, it is common to restrict attention to cost functionals that are linear combinations of fixed basis functions. The usual approach is to iterate over the weights, numerically solving an optimal control problem at each iteration to compare predicted and observed trajectories. Our approach is to construct an augmented system model based on the maximum principle, in which the costate is a latent variable, and to recover the weights using nonlinear state estimation. We will show that our approach significantly reduces computation time and often produces better results as well, particularly when the corresponding optimal control problem has many local minima. We will consider several applications of recent interest, such as the characterization of human walking paths and the implications of this result to planning and control of humanoid robots.

Biography:

I did my first two years of undergraduate studies in France at the Skema, and transferred to Virginia Tech in 2008 to complete my B.S. in Aerospace Engineering. I joined the RMS Lab at the University of Illinois at Urbana-Champaign last summer as a Masters student in Aerospace Engineering. My interests are in optimal control, neuroscience and brain-machine interfaces.

_______________________________________________________________________________________________________________