ACTIVE PIXEL SENSOR FOR A POLARIZATION-DIFFERENCE CAMERA

Advisors: Dr. Nader Engheta and Dr. Jan Van der Spiegel

Polarization-sensitive vision is well documented as serving in navigation for many animals, but some types of biological polarization-sensitive vision may enhance object visibility in scattering media. Because neither the human eye nor conventional cameras are polarization-sensitive, artificial polarization vision systems must be designed to exploit the polarization of light; artificial polarization-difference imaging has been shown to be capable of enhancing target detection in scattering media. Previous polarization-sensitive cameras required external processing, were not real-time, and used relatively large amounts of power. A CMOS active pixel sensor is presented for use in a low power, portable, real-time polarization-difference camera. Pixels were designed for integration with a diffractive optical element polarization analyzer. Column readout circuits include fixed pattern noise suppression. In addition, a scaling methodology to enhance system performance and to correct for non-ideal polarization analyzers is presented. View Paper | View Slides

ANALYSIS AND DESIGN OF SUBWAVELENGTH DIFFRACTIVE OPTICAL ELEMENTS FOR POLARIZATION-DIFFERENCE IMAGING

Advisors: Dr. Nader Engheta, University of Pennsylvania; Dr. Dennis Prather, University of Delaware

Polarization-difference imaging (PDI) has been proven to significantly enhance the detection of targets and their surface features in scattering media. We are currently building a portable prototype PDI camera that can capture a scene at orthogonal linear polarizations and can produce PD images in real time. This paper details the design and analysis of subwavelength diffractive optical elements (DOEs) that will function as linear polarization analyzers in the PDI camera. Using Matlab software, we derived pulse-width modulated subwavelength features from continuous phase functions and modeled the wave propagation through the components using the finite-difference time-domain (FDTD) method. Various lenses operating at incremental wavelengths of the visible spectrum were designed, analyzed, and evaluated. DOEs designed for 400, 450, and 500 nm incident waves performed the most desirably; their intensity ratios, obtained by dividing the focal plane’s maximum electric field intensity from TM wave illumination by that from TE wave illumination, remained relatively constant at approximately 2.5:1. View Paper

MICROFABRICATION OF HETEROGENEOUS, OPTIMIZED COMPLIANT MECHANISMS

Advisor: G. K. Ananthasuresh

Metal deposition by electroplating has been extensively used and analyzed for industrial applications, but it remains insufficiently characterized for fabricating compliant micromechanisms. This research presents a new fabrication process for heterogeneous optimized compliant mechanisms that consists of placing two materials side by side as opposed to the convention of stacking materials layer by layer. The fabrication process uses standard bulk micromachining and electroplating techniques. This paper reviews and discusses the theory and method of electroplating for obtaining optimum structural and surface morphology. We also present a novel method for protecting a silicon wafer from wet etching using black wax. In fabricating these two-material mechanisms, we found that optimum microstructures can be obtained by electroplating at a low-current density while maximizing the reaction kinetics. View Paper | View Slides

FINGER MOUSE AND TEXT-TO-SPEECH APPLICATION AS ADDITIONS TO THE SMART WHEELCHAIR

Advisor: Dr. Jim Ostrowski

The smart wheelchair project is a unique investigation into the possibilities of helping the impaired navigate in a mobile chair. Many disabled people who need the help of a wheelchair to move about also need help communicating orally. My project allows the “walking” wheelchair to do some “talking.” I developed the addition of a communication program for the wheelchair as well as a finger mouse to be implemented into all the programs. The finger mouse is a switch button small enough to wear on one’s finger. When the button is pressed, a signal is sent out from the transmitter and picked up by the receiver, which sends a signal through the parallel port of the computer to execute the desired application. The application I created is a speech program that speaks text messages. The finger mouse and speech application are connected through a communication display interface. The mouse, display, and speech software work together by speaking phrases when the mouse is clicked over the icon linked to that phrase. Thus the communication program gives the freedom of speech to anyone using the wheelchair for free range of motion. View Paper | View Slides

ROBOTIC BUTLER: PHASE I

Advisors: Dr. Vijay Kumar and Dr. James Ostrowsi

ABSTRACT: The contents of this report describe the specifics of a project aimed to program a “domestic” robot to follow an individual or specified target using sensor technology. Ultimately, the target will be, though not restricted to, the “Smart Chair,” a wheelchair that facilitates physically handicapped citizens with speaking and maneuvering. This would be of significant assistance to those with physical handicaps in the sense that the robot would follow them around to aid in any area possible.

The domestic robot referred to in this project is called the Cye sr. Although the Cye sr comes with its preprogrammed Map-N-Zap software, the decision was made to modify its source code to program the robot directly. The original prototype used one IR receiver and one IR transmitter. After testing, improvements were made involving the addition of only another IR receiver, which required the use of a microcontroller. Upon completion of the project during the allotted period, the goal was generally achieved as the robot essentially follows a target using an infrared sensor. More testing needs to be conducted for further improvements. View Paper

NEURAL NETWORK WITH APPLICATION TO REAL TIME PHONEME RECOGNITION

Advisors: Dr. Paul Mueller, Dr. Jan Van der Spiegel

In hopes of improving real-time speech recognition, a biologically based phoneme recognition algorithm was implemented on the NP-4 neural computer. The NP-4 neural computer, which contains programmable interconnects, neurons, synapses, and synaptic time constants, is extremely useful in computation of real-world dynamic patterns as they occur in speech. Prior to this summer, some implementations were done on the NP-4 neural computer, and the goal during this summer was to improve work done in early stages. The newly developed algorithm, implemented in the host computer, allows neurons to be trained to respond to a particular phoneme. Testing was performed once the network was trained to find the overall responses of the neurons. The algorithm shows much promise for recognition of phonemes, with over 90% positive responses. View Paper | View Slides

EXPERIMENTALLY DETERMINING THE INTERNAL FRICTION OF CONDUCTIVE FIBERS

Advisor: Dr. J. J. Santiago-Avilés

Conductive polymer fibers have the potential to take electronics to the molecular scale by allowing current to pass along a single molecule. Unfortunately, their use has been limited by difficulty in determining their mechanical properties. It is the goal of this research to develop a technique for measuring the internal friction of conductive polymer fibers. Unfortunately, I was not able to use a conducting polymer for this research, but instead used gold wire. I forced the wire into harmonic resonance by exposing it to an oscillating magnetic field while passing current along the wire. Using a phototransistor and LED I was able to observe the decay of the fiber from resonance in air. Were this technique used in a vacuum, it would be possible to determine the internal decay of the sample. View Paper | View Slides

UNMANNED BLIMP

Advisors: Dr. J.P. Ostrowski and Dr. J.J. Santiago

For the 2001 SUNFEST Research Program I am using sensor fusion of GPS, and rate gyroscope systems to automate the flight control and measure the dynamics of a 30 foot unmanned blimp. This project is parented by Professor James Ostrowski of the University of Pennsylvania GRASP Lab. Using an onboard computer, a Garmin GPS36 GPS sensor, a pair of rate gyroscopes with supporting hardware, I developed a custom operating system in Java as well as sensor integration software for this operating system. I have developed a system to measure blimp dynamics while it flies and created a set of mathematical models to describe the blimp. This project is sponsored by the National Science Foundation, the University of Pennsylvania School of Engineering and Applied Sciences, the University of Pennsylvania Science and Technology Wing living-learning program, and the University of Pennsylvania College House System. View Paper | View Slides

USING IMPROVED OBJECT DETECTION IN ROBOCUP SOCCER FOR COLLISION AVOIDANCE AND RECOVERY

Advisor: Professor Jim Ostrowski

RoboCup legged soccer is an international soccer competition between autonomous, four-legged robots. In past competitions, collisions between robots were common. My objective, therefore, was to augment the robots’ sensor systems to allow them to avoid and recover from such collisions; this would both speed up the game and enable the robots to dribble the ball around other players. I have increased the accuracy and stability of the ball’s position estimation and have added a calculation of the ball’s relative velocity. Using this velocity, I have developed an algorithm to determine if a robot has collided while trying to get to the ball; if the velocity remains near zero for too long, the robot will try another path to the ball. I have also greatly increased the accuracy of the detection of other robots and have added the ability to store the positions of all other players on the field. A dribbling robot will then use these position estimates for collision avoidance, using the method of artificial potential fields. I describe a number of modifications that were necessary for this method to work well. Finally, for both the ball and other players, I have incorporated readings from the robots’ infrared distance sensors, and have added an update of the relative position of an object based on the observer’s own movements. View Paper | View Slides

FABRICATION AND PROPERTIES OF RAPID THERMALLY PROCESSED CARBON NANOFIBERS

Advisors: Dr. Jorge J. Santiago-Aviles, Yu Wang

The fabrication and properties of carbon nanofibers are interesting because of the fiber’s expected conductivity properties. If we can successfully control the fabrication and manipulation of carbon nanofibers we should be able to build nanoscale electronics. In this attempt, we have been able to successfully control some of the properties for the fabrication of nanofibers. Carbon nanofibers were obtained by electrospinning a polymer solution, polyacrylonitrile, and then heating it on a vacuum furnace. Thermal analysis helped us determine the minimum heating temperature that the fibers must be exposed to obtain carbon fibers. The relationship between voltage, and the amount of solvent, with the thickness of the fibers was studied using a scanning electron microscope. A strongly non-linear relationship between the voltage and the thickness suggests a sinusoidal relationship. On the other hand, a linear relationship between the amount of solvent and the inverse of the thickness was observed. The resistivity of a single carbon fiber was studied; its total resistance agreed with the sum of the resistances of smaller sections. This implies that the resistance along the carbon fiber is linear. Raman microscopy established the chemical bonding of the carbon nanofibers as a function of the heating temperature. View Paper

SPIN-POLARIZED ELECTRON TRANSPORT IN SINGLE-WALLED CARBON NANOTUBES

Advisor: Dr. A.T. Johnson, Jr.

A mounting research effort geared towards miniaturization of electronic devices has led to the emergence of a new field called molecular electronics. A particular type of carbon molecule called the nanotube has been playing an integral role in steering this revolution. Nanotubes are micron-long and nanometer-thick cylindrical shells of carbon that have been found to have excellent electrical and thermal conductivities. Their ability to behave as metals or semiconductors depending on the tube structure has led to their incorporation in nano-scale circuits as diodes, field effect transistors and quantum wires. Conventional electronic devices have exploited the charge of electrons to generate current but there has also been a burgeoning interest in harnessing the intrinsic spin of electrons to produce spin-polarized current- a field that has come to be known as spintronics. My project tries to investigate the transport of spin-polarized electrons through single-walled carbon nanotubes. In a broad sense, my project deals with not only extending contemporary macroscopic technology to the molecular level but also exploring a new conduction mechanism to evaluate the potential of nanotubes as nano-spintronic devices. View Paper | View Slides

ENERGY DISSIPATION IN CONDUCTIVE POLYMERICFIBER BUNDLES: SIMULATION EFFORT

Advisor: Jorge Santiago-Avilés

This work built a mathematical model and simulation scheme to support a physical experiment on energy dissipation from vibrating conductive polymer fibers. The fundamental idea in the physical experiment is to oscillate a magnetic field from a coil, so a conductive polymeric fiber with a current through it, is excited to oscillations by the resulting Lorentz force. I used Matlab software to graphically represent the functional dependence of the vibrational amplitude on force amplitude, frequency, and quality factor. The model is able to develop three two-dimensional and three three-dimensional graphs. The model lets the user choose a graph and either use the default values or adjust the values of the constants and variables for a particular situation. View Paper | View Slides