Seattle WA, USA
tfolsom at ieee dot org; 206-522-6441
Dr. Folsom specializes in the areas of Computational Intelligence, Embedded Systems and Machine Vision. He has 6 years of teaching and 19 years of engineering and programming experience and is a licensed Professional Engineer in the state of Washington. He has designed both software and hardware for industrial microcomputer inspection and control systems and has used neural network and standard image processing techniques for machine vision systems. Dr. Folsom has published over 35 papers. He has bicycled around the world (1976-1978) and been a visiting professor at Northwest University, Xian, China (2004). He speaks fluent French and several other languages.
University of British Columbia (www.ubc.ca, www.ubcthunderbird.com) Vancouver, BC, Canada (2006-present). Team leader for the DARPA Urban Challenge robot race.
DigiPen Institute of Technology, (www.digipen.edu) Redmond WA (2000-2006) Full Professor. DigiPen offers baccalaureate and masters degrees in game programming and game art. It is the only school in the world to have consistently placed two or more student projects in the Independent Game Developer Conference’s annual student showcase. Classes taught:
CS 200, 250 Computer Graphics I, II
CS 240 Selected Topics: Artificial Intelligence
CS 315 Low Level Programming
CS 365 Software Engineering
CS 370 Image Processing
CS 380 Robotic Intelligence
CS 381 Machine Learning
CS 420 Graphics File Formats and Data Compression Techniques
EE 210, 260 Digital Electronics I, II
Quest Integrated Inc. (www.qi2.com) Kent, WA (1984-1990) Senior software engineer; (1994-2000) Project manager. Also known as Flow Industries or Flow Research. A small company doing contract research and development for public and private clients. Quest did much work under the federal Small Business Innovation Research (SBIR) program and was the leading company for winning awards in some years.
Numerous R&D projects for various clients. Most involved programming an embedded system in C or C++ and custom hardware.
Macotech Corp., Seattle, WA (1980-1984) Software Engineer. Macotech manufactured adaptive control systems for industrial milling and drilling machines. No longer in business.
Wrote microcontroller firmware in assembly for real-time control of large milling and drilling machines. Designed hardware and software for paper-tape punch. Maintained and modified Fortran code to predict metal cutting behavior.
Sperry Univac (www.unisys.com), Goddard Space Center, Greenbelt, MD. (1973-1976) Associate scientific programmer. At the time, Univac was the world’s 3rd largest computer company.
As a member of a team of 12 people, wrote the real-time operating systems for two unmanned spacecraft as a NASA subcontract. All code was in assembly and he operated the main-frame computer and peripherals to debug
Senior member of IEEE.
Member of Toastmasters at the level Able Toastmaster (silver).
Selected for Who’s Who in America, 2007 (www.marquiswhoswho.com)
Currently Team Leader for Thunderbird Robotics. This is the attempt by the University of British Columbia to win the DARPA Urban Challenge. (2006-2007)
Snowstorm vehicle with Reuven Granot, Robert Hecht-Nielsen, John Meech, Tyler Folsom
Dr. Folsom served as a consultant to Team Sleipnir, which was an attempt to win the DARPA Grand Challenge. Vision was the primary obstacle detection system. Dr. Folsom wrote the software for edge detection and binocular stereo matching. (2005)
Dr. Folsom managed this Air Force contract, which went through several phases for a total of $1.1M. Dr. Folsom was the primary liaison with the customer and wrote proposals that increased funding from the initial $100,000. He directed a small team, designed the system, and wrote code in C++. The project used artificial intelligence techniques to automate the task of formatting flight test data while avoiding combinatorial explosion. A software architecture was designed to allow integration into software from different vendors with changes limited to the modules that interface to the database and hardware. The system was implemented in a form compatible with the TIMS software provided by Data General. A commercial grade system using Microsoft COM and interfacing to various telemetry databases was produced. Runs on Microsoft Windows™ and Unix. (1998-2000).
Inspection of Composite Material
Dr Folsom was principal investigator for research performed for the U.S. Air Force Wright Laboratory. Composite material is often constructed on expensive machines that lay down a layer of tows. If these ribbons of material are too far apart or too close together, the part may be defective. It is easy to remove an imperfect layer and replace it, but if a bad layer is allowed to persist it can jeopardize the integrity of a part that requires hours to lay up. A prototype optical instrument using laser shadowing was developed to measure the gap widths. (1995)
QUEST built two systems to measure critical dimensions of internally and externally threaded parts. The work was performed under Phase II SBIR contracts for the Army and Department of Commerce. Dr. Folsom was project manager. Each system consists of Microsoft Windows™ software and a laser optical-triangulation scanner mounted on a precisely controlled three-axis translation and rotation stage. The Windows software interacts with a DSP board to acquire thousands of data points. The data is then fit to a 3-D model with the results displayed both graphically and as numerical measurements. (1997)
Determination of a Machining Path to Cut Honeycomb Material
In this commercial contract, QUEST developed a prototype automated flexible abrasive-waterjet (AWJ) cutting center for precision slot cutting in manufactured parts. Dr. Folsom used the vision capabilities of an Adept robot controller to develop a tool that rapidly maps the locations of hundreds of individual small cells in each part and transfers the data to an AWJ cutting station. The data are used to direct the AWJ beam so that slots are cut precisely (within ±0.005 in.) down the center of irregular cell rows at high nozzle traverse speeds. (1994.)
Neural Networks Modeling Cortical Cells for Machine Vision
This doctoral dissertation was a study of wavelet transforms and biological visual systems. It has produced C/C++ language computer software. The research was supported by a grant from the NSF Neuro Engineering Program based on a proposal that Mr. Folsom wrote with his major professor. (1994)
Neural Networks in Machine Vision
Dr. Folsom was the principal investigator for a research award from the National Science Foundation. This project was entitled “Machine Vision for Composites Manufacturing,” and used a modular neural network approach to find edges in images of laminated carbon graphite material. Conventional image processing techniques do not perform satisfactorily since this material is highly specular, nonisotropic, and reflects poorly. The project also evaluated the feasibility of making VLSI analog chips to perform this inspection task. (1989)
Last Updated: August 27, 2007