Clinic Projects

Please click on a link below to view the Harvey Mudd College Computer Science Clinic projects for the corresponding time period.

Clinic Projects for 2003-2004

Developing A Personal Digital Transcriber

Client
CPI Corp.

Faculty Advisor
Professor Elizabeth Sweedyk

Student Team
Matt Livianu (Project Manager), Melissa Federowicz, Matt Ferlo, Colleen Hamilton
Computer Product Introductions (CPI) is interested in the development of a portable real-time speaker-independent speech-to-phoneme system which transcribes, compresses, stores, and plays back sound files. Charged with a subset of these goals, the CPI Clinic team has been asked to research speech-to-text software and phoneme readability, and to modify a batch speech-to-phoneme software in order to produce readable annotated phoneme output.

Network Security Controller Performance Validation

Client
Cryptek, Inc.

Faculty Advisor
Professor Geoff Kuenning

Student Team
Keith Stevens (Project Manager), Matthew Beaumont-Gay, Victoria Krafft, Alex Popkin
Cryptek, Inc. provides secure communications products for the government and private sectors. Our project is to stress-test a central component of the Cryptek system, the Network Security Controller (NSC), by creating software that mimics the encrypted network traffic from thousands of hardware nodes known as DiamondLinks. We are developing software that will go through an authentication protocol simultaneously for a large number of simulated DiamondLinks, and record the NSC's performance.

Document Imaging Using Off-the-Shelf Digital Cameras

Client
Laserfiche

Faculty Advisor
Professor Zachary Dodds

Student Team
Ed Heaney (Project Manager), Zak Andree, Zach Clegg, James Darpinian
The goal of our Clinic project is to develop a software module that converts a digital photo of a document into an image that looks scanned. It finds, straightens, and orients the document while providing confidence values to ensure reliability. Also featured are methods to help reduce the effects of distortion and lighting irregularities introduced by the camera.

Diabetes Data Management Software API Design and Development

Client
Medtronic/MiniMed

Faculty Advisor
Professor Belinda Thom

Student Team
Jessica Fisher (Project Manager), Aja Hammerly, Mark Fredrickson, Jon Huang
With approximately 17 million people in the US with diabetes, Medtronic MiniMed has produced several distinct lines of diabetes devices to aid in the treatment of the disease. These devices, however, do not utilize a standard communication format. The Clinic team is designing and implementing an extensible interface that will unify communication with Medtronic MiniMed's current and future insulin pumps, glucose sensors, and related diabetes technology.

Automated Management of Community Websites

Client
NeonGecko.com, Inc.

Faculty Advisor
Professor Melissa O'Neill

Student Team
Drew d'Avis (Project Manager), Brian Merdian, Mark Nelson, Jenny Xu
This project developed tools to analyze posts made to NeonGecko's on-line discussion forums, classify posts into known topic areas, and identify new topics. The team applied supervised learning techniques (trained with precategorized posts) to determine topics of new posts. The team used clustering techniques to identify groups of posts that are markedly different from existing topics and thus might be new topics. Project work stressed extensive technique testing and delivery of a functional system.

Anomaly Detection in Health and Status Telemetry Data

Client
Northrop Grumman Corporation

Faculty Advisor
Professor Robert Keller

Student Team
Erika Rice (Project Manager), Daniel Marley, Gabriel Neer, Jesse Ruderman
Detection of anomalies in satellite health and status data requires real-time processing capabilities in order to reduce the ill effects of equipment malfunctions and other undesirable events. The team designed and implemented an extensible software architecture that enables anomalies to be detected and displayed in visual form. In addition to preset monitoring capabilities, our system provides learning capabilities based upon techniques from adaptive signal processing and adaptive resonance theory.

Patient Alignment Through X-Ray Imaging

Client
Optivus Technology, Inc.

Faculty Advisor
Professor Patrick Little, Professor Elizabeth Sweedyk

Student Team
Michael Tuck-Lee (Project Manager), Mjumbe Poe, Knut Strom-Jensen, Anand Vemuri
The goal of this Clinic project was to upgrade the alignment system on Optivus Technology's Proton Beam Therapy System (PBTS) Eye-Beam Line by replacing the obsolete components with digital imaging devices and an accompanying software package. This involves both mechanical and electrical engineering to implement the hardware requirements, and software development to create the supporting software.

Intrusion Detection and Prevention Based on Immunology

Client
Raytheon / BioStar Group

Faculty Advisor
Professor Mike Erlinger

Student Team
Robert Bailey (Project Manager), Ian Ferrel, Kevin Pang, Jeff Scherpelz
The goals of this Clinic project are to research, design, and develop a host-based Artificial Immune System (AIS) for intrusion detection based on the biological immune system. The main areas of focus for the project are to investigate and improve the idea of "self" in an AIS for more accurate anomaly detection, and to allow for a dynamic concept of "self" without compromising security.

Investigating Artificial Immune Systems

Client
Raytheon

Faculty Advisor
Professor Mike Erlinger

Student Team
Brandt Erickson, George Kuan, Michael Terkowitz, Andrew Yip
Conventional methods of computer security have many shortcomings. The team focused on the application of biological immune system paradigms to computer security. After assessing the various aspects of the biological immune system, as well as the current state of artificial immune system research, the team designed and implemented an extension to an existing artificial immune system.

Wireless-Assisted GPS Applications

Client
SnapTrack (A QUALCOMM Company)

Faculty Advisor
Professor Christopher Stone

Student Team
Corey O'Connor (Project Manager), Alice Liu, Stuart Mershon, Tatsuya Oiye
The SnapTrack Clinic team developed two cell-phone applications, Direction Finder and Friendar, which demonstrate SnapTrack's Assisted-GPS technology. Direction Finder demonstrates the utility of a cell phone user being able to quickly acquire driving directions from their current location. Friendar demonstrates the fast update speed and high accuracy of A-GPS by allowing users quickly to pinpoint their friends' locations.

Launch Range Countdown Clocks

Client
The Aerospace Corporation

Faculty Advisor
Professor Ran Libeskind-Hadas

Student Team
Jonah Cohen (Project Manager, fall), Joshua Smallman (Project Manager, spring), Amanda Parmelee, Alan Strohm
Countdown clocks, a common tool of launch ranges, are used to synchronize and control the numerous and complex series of actions leading to the launch of a space vehicle or guided missile. However, countdown clocks rely on a standard for time distribution and synchronization that, in comparison to modern digital protocols, is anachronistic and needlessly restrictive. The Clinic team will present an entirely new standard for the management of range countdown clocks founded on modern and effective protocols, such as the Network Time Protocol (NTP) and the Hyper Text Transport Protocol (HTTP), which will improve both the accuracy and flexibility of countdown time services.