Advanced Teleoperation Control System for PR2 Humanoid Robot

Advisor: Dr. Camillo J. Taylor

Abstract: The ultimate achievement in robotics is to build a robot that can perform perform a range of tasks by human. One such robot that is getting closer to this ability is the PR2 humanoid robot, from Willow Garage. The PR2 has capabilities that allow the user to program different movements in order to achieve multiple assignments. Being an open source robot, the PR2 allows users to read previous codes and manipulate them in order to complete new tasks. The purpose of this research is to output what the PR2 sees to the Head Mounted Display (HMD) worn by the user, and control the PR2’s head with a motion tracking system (Vicon). The Vicon system captures the movement of markers placed along the HMD, and outputs an array of numbers that will go through a Python code in order to retrieve the necessary coordinates. These coordinates are sent to the PR2 which will enable it to move the head to the specific point in space. With all the components working together the user is capable of manipulating the PR2’s sight for his own use. View Paper | View Slides

NANOELECTRONIC SENSOR FOR DETECTION OF PROSTATE CANCER BIOMAKERS

Honorable Mention

Advisor: Dr. A.T Johnson

Abstract: Prostate cancer is the second cause of cancer death in American men; it is known that early detection of the disease is one of the most important tools for successful treatments. Unfortunately, current methods of diagnosis are either invasive or require high concentrations of prostate cancer biomarkers in order to detect the disease accurately. Here, we present an improved method for early detection, a nano electronic sensor for prostate cancer. This device relies on electrical sensing using a field effect transistor where the semiconducting channel is a functionalized single-walled carbon nanotube. These novel carbon structures have unique electronic and chemical properties that allowed us to fabricate a highly specific and highly sensitive nano sensor. Moreover, we designed a channel that allowed us to controllably deliver small amount of samples to the devices. This will allow real time detection since electrical measurements can be taken at same time the device is exposed to the biomarker. It is believed that this project could set up the platform to eventually have an array of these devices to detect several biomarkers in one step. View Paper | View Slides

Characterization and Design of Organic Field-Effect Transistor Circuits for Sensing Bioelectromagnetism

Winner of the Best Sunfest Award

Advisor: Cherie Kagan

Abstract: Current scanning technology for the brain and heart requires electrodes to be placed on the surface, with a wire connected to each electrode. Because the electrodes are large, it is impossible to achieve a high sampling resolution of the signals from the tissue being scanned. Silicon based structures are not very suitable for this application. They have a rigid planar surface which prevents them from capturing a high degree of information from the three dimensional structure of the brain or the heart. However, an organic transistor, fabricated on flexible plastics, should be able to conform to and output a high degree of information from a three-dimensional structure. To test the ability of organic transistors to read data, an organic field-effect transistor was placed in a common source configuration. This configuration, when used with an organic transistor, allows the electrodes to conform to the structure, appear in higher density, and cover a larger area. A 50µm channel length transistor was able to operate with a low frequency gain of 0.97V/V and with a cutoff frequency of 91Hz. Organic field-effect transistors fabricated into circuits that are more complex were analyzed so that they could be considered for amplification purposes. A 6µm channel length inverter showed a low frequency gain of 3.2V/V and a cutoff frequency of 145Hz. A 10µm channel length cascode showed a low frequency gain of 2V/V with a cutoff frequency of 220Hz. These results suggest that organic field-effect transistors have the ability to measure and amplify a small signal and become an effective tool for mapping high-density signals of the brain and the heart. View Paper | View Slides

IN VITRO INVESTIGATION OF CYTOKINE-MEDIATED NUCLEUS PULPOSUS DEGENERATION

Advisor: Dawn Elliott
Post-Doctoral Fellow: Lachlan Smith

Abstract: Degeneration of the lumbar intervertebral discs is strongly implicated as a cause of low back pain, and may also lead to impaired mobility. A lack of understanding of the pathomechanisms that underlie degeneration limits our ability to develop biological treatments that both alleviate painful symptoms and restore function. The process of degeneration is characterized by up-regulation of pro-inflammatory cytokines—particularly interleukin 1 beta (IL1β) and tumor necrosis factor alpha (TNFα)—within the central nucleus pulposus. Furthermore, the increased production of these cytokines is not matched by increasing amounts of their inhibitory regulators, resulting in an imbalance of catabolic and anabolic activity. In this study, we developed an in-vitro model of the nucleus pulposus that was used to investigate the effects of IL1β and TNFα on composition and mechanical function. In addition, we examined the capacity of IL1 receptor antagonist (IL1ra) and soluble TNF receptor 1 (sTNFR1), inhibitors of IL1β and TNFα respectively, to mitigate cytokine-mediated functional and compositional changes. Our results demonstrated that short-term exposure to IL1β, but not TNFα, causes loss of matrix components that significantly compromises mechanical function, suggesting that IL1β plays a more direct role than TNFα in driving matrix degradation in the nucleus pulposus. Our results also demonstrated that IL1ra can effectively prevent compositional and functional changes induced by IL1β, highlighting its therapeutic potential. View Paper | View Slides

Autonomous Laser Locking System

Advisor: Professor Daniel D. Lee

Abstract: The purpose of this project is to create an autonomous laser locking system for a robot that will be used in the Multi Autonomous Ground-robotic International Challenge (MAGIC). Locking onto a target has many applications in robotics. The laser could be replaced by a flashlight, a camera, a projectile launcher, etc. Solving this problem requires the use of coordinate transformation matrices to deal with multiple reference frames. It also requires sensor analysis to determine positions of both the robot and the target. In this paper I describe an approach to solving this problem. View Paper | View Slides

Pediatric Dynamometer Using Piezoresistance Sensor

Honorable Mention

Advisor: Dr. Jay N. Zemel

Abstract: External forces on the body have long been known to have a large effect on children’s bone growth and development. Researchers at Children’s Hospital of Philadelphia (CHOP) want to explore this relationship. However, there is no current technology that directly measures the forces applied to the body. Dr. Babette Zemel, from CHOP, and Dr. Jay Zemel, from ESE at Univeristy of Pennsylvania, have been developing an in-shoe physical activity dynamometer (FootPAD), which will directly measure forces felt through children’s feet. The past versions of this device have been developed using piezoelectric sensors; however, the drift caused by temperature changes in the shoe was unacceptable. This study looked into using piezoresistance sensors in the device instead of the piezoelectric ones. Preliminary tests with the sensors showed that they did not have the same temperature problem and that the sensors could accurately measure changes in force within the 10% accuracy needed. A circuit for the device using these sensors was designed and built, however further work with the software is needed before the device can be fully implemented and tested. View Paper | View Slides

NAVIGATIONAL SENSING FOR THE EDUBOT

Advisors: Dr. Daniel E. Koditschek , Dr. Galen Clark Haynes

Abstract: Automation – making a robot perform human functions but without human control – is the overarching goal of robotics. An important aspect of autonomous robots is the ability to detect and avoid obstacles. The goal of this project was to determine the feasibility of enabling a hexapedal robot, known as the EduBot, to generate maps of its surrounding area. This was accomplished by equipping the EduBot with a laser scanner, which was then controlled by the ROS robotic software package. The laser scanner was used to measure the distances to obstacles surrounding the EduBot. These measurements were then used by a ROS program to generate a map of the environment. The EduBot mapped various areas within a building to provide a large sample for analysis. These maps were then compared with the actual area to determine their accuracy. Once software and hardware errors discovered in early maps were corrected, the EduBot was able to generate accurate representations of its surrounding environment, thus allowing the EduBot to sense it surroundings. This suggest that autonomous navigation by the EduBot is possible. View Paper | View Slides

Evaluation of Composite Electronic Materials Based on Poly (3, 4 – propylenedioxythiophene/Poly – (p – Naptheleneethynylene) Wrapped Single Wall Carbon Nanotubes for Supercapacitors

Advisor: Jorge Santiago

Abstract: Supercapacitors have emerged to do major advances in energy storage and technology. This paper summarizes the performance data of a specially designed poly (3, 4 – propylenedioxythiophene) PProDot based conducting polymer for use in a p-type supercapacitor. Performance data of these polymer composite electrodes are also compared with those of a poly (3, 4 – propylenedioxythiophene/poly – (p – naptheleneethynylene) wrapped single wall carbon nanotube supercapacitor. Longevity of the system was analyzed to determine life span and durability of both PProDot and PProDot/PNES/SWNT based devices. Both composites were characterized using Scanning electron microscopy (SEM). View Paper | View Slides

MATERIAL TESTING OF SHAPE MEMORY POLYMERS FOR MODULAR ROBOTICS APPLICATIONS AND DEVELOPMENT OF A PROTOTYPE SMP GRIPPER FOR mini-PR2 ROBOT.

Advisors: Paul White and Dr. Mark Yim

ABSTRACT: Goals that have yet to have been realized in the field of modular robotics include low cost mass production, and scalability. Shape memory polymers are light weight, low-cost, and have a large degree of flexibility in material design. For these reasons these polymers have the potential to help reach current goals of modular robotics. The percent recovery and force of these materials for use in actuation was tested in this research. A prototype SMP gripper consisting of two reconfigurable one way SMP actuators was developed that successfully picked up an object weighing one gram in 90% of trials. This SMP gripper is compatible with the CKbot modules from Dr. Mark Yim’s laboratory at the University of Pennsylvania and it was installed on an existing modular robot, the mini-PR2, and the robot was programmed to use the gripper successfully. While there remain other obstacles in scaling and mass production problems, the shape memory component of this gripper could be produced on a wide range of scales, and with the proper equipment, made in large quantities for mass production. Future work to improve the SMP gripper will need to involve developing a method of making the gripper reversible. View Paper | View Slides

Reducing Anchor Loss in AlN Contour Mode Resonators

Advisor: Gianluca Piazza

Abstract: Many researchers have devoted money and effort toward developing of MEMS resonators. Much of this effort has been devoted to increasing the quality factor, a measure of the energy lost by the resonator. Fabricating new resonators for testing is costly and time consuming. Methods for accurately simulating the quality factor of MEMS resonators would increase development efficiency. By design, the resonator body is attached to the substrate via tethers. Due to the immense size of the substrate with respect to the resonator, any energy traveling from the resonator to the substrate is lost. In order to accurately simulate the resonator it is necessary to have a semi-infinite domain that behaves like the substrate. The perfectly matched layer (PML) feature of COMSOL FEM software serves as an artificial medium: surrounding the smaller simulation-substrate and absorbing any radiation escaping from the resonator. Before resonator designs can be simulated, though, it must be confirmed that the PML does not affect the quality factor or resonant frequency of the resonator. In order to determine the conditions for accurate simulations a resonator is modeled in COMSOL and simulations are performed while varying the PML width, maximum mesh element size in the PML sub-domain, substrate size, maximum mesh element size in the substrate sub-domain, and tether location. The quality factor is determined by calculating the admittance and using 3dB bandwidth. The predicted results are that the quality factor will decrease with mesh density and will converge around 5-micron mesh size. These results will allow accurate simulation of tether designs meant to reduce anchor loss, the biggest problem facing the development of MEMS resonators. View Paper | View Slides