Summary
I joined SVS when it was 34 people strong in 1995. They grew to over 100 employees before Boeing bought the company. During my tenure at Boeing-SVS I was involved in many different projects, mostly related to directed energy programs. Most of the folks at SVS came from the Air Force and had participated in projects like the Airborne Laser Lab and other 'star wars' endeavors.
What I particularly enjoyed about SVS was the diverse set of projects that I was able to work on. These included building the prototype trackers for the Airborne Laser, the recording system for the Advanced Tactical Laser, a tracking demonstrator for the C-RAM program, a jet fuel reclamation station, a video-based tester for the Lightweight Exoatmospheric Projectile, and an internal R&D project for a solid-state laser.
Accomplishments
Senior Embedded Software Engineer/Real-Time Software Engineer
- Team lead for the Airborne Tactical Laser (ATL) pointing and tracking embedded software. Duties included requirements development and management, design and development of the digital recording station, system integration and test. Led team to successful delivery of the system to the customer. Worked alongside mechanical, electrical, controls and optical engineers during all phases of development.
- Team lead for a counter-mortar pointing and tracking system (CRAM). Working with a controls engineer, we rapid-prototyped the software for gimbal control that would automatically acquire and track mortars from a radar hand-off. Supported field tests in Yuma on multiple occasions; during the first visit, our system beat out all pre-existing competitors and became the gold standard for position and rate of a launched mortar. Transitioned technology to an internal R&D system that supported other field tests, including tracking ground to air missiles.
- Performed initial embedded systems design of an internal R&D LADAR (Laser Detection and Ranging) program. System captures return signals from a highly focused LADAR at GHz rates and stores the data for post-processing and analysis.
- System architect for the embedded and application-level control systems for an internal R&D high energy laser. Led a small team in early stages of the embedded system development, including requirements definition and management, system design, and integration with mechanical and optical components. Drove interchanges with heads of other disciplines to define system level requirements and interface specifications.
- Helped define and implement agile development practices; process definitions were adaptable to both large and small projects. Introduction of the practices reduced the time for subsequent development and test iterations for the ATL program.
- Developed first generation of the Airborne Laser (ABL) tracking systems which contained five separate optical tracking systems. One such tracker was fed by a 5KHz 128x128 CCD camera and incorporated 32 PPC processors.
- Created a power line tracking system which resided on a helicopter. Acquired imagery was used to detect issues with wires and insulators.
- Designed and developed the tracking system for the Lightweight Exo-Atmospheric Projectile (LEAP) test-bed. The test-bed was for simulating an air-to-air missile and was instrumental in improving the accuracy and optical target acquisition capabilities of the LEAP system.
- Built software to control an environmental pumping station used to remove oil and gas from underground tanks and spills. Systems were installed on Kirtland and Hickam Air Force Bases.
- Re-wrote software to fix faulty telemetry acquisition system for the High Altitude Balloon Experiment (HABE). Final product worked consistently and accurately for all subsequent lab tests.
- Earned SVS Performance Award; 1997.
Highlights
Airborne Laser
I was software lead for the pointing and tracking system on the Airborne Laser. As part of my responsibilities, I was the prime developer on the video trackers. The most demanding of the video trackers had a source CCD that ran at 5 KHz and had eight data outputs. We used 32 PPC processors to collect and process each frame. Each PPC reported data in a tree-format; the root processor determined the final track error and reported it to the optical controls. There were 5 other trackers, some for target detection and some for laser guidance. For development and test, I created a series of user interfaces in Tcl/Tk for command/control of each tracker.
C-RAM
I really enjoyed this project because I got to see things blow up from the results of my labor (I've also seen PCBs ignite, which is quite a different adventure). Watching a mortar explode from a rapid-fire phalanx at night is quite spectacular! A small group from SVS rapid-prototyped a mortar tracking system and deployed it to the C-RAM tests. The system primarily comprised of a gimbal and a single IR camera. The controls accepted a radar signal for the physical position of the mortar; once the target was in the field of view, the system tracked the mortar until it was either decimated or to impact. Our tracking algoirhtms worked in both day and night scenarios.
The software ran on a PPC running VxWorks. Most of the control algorithms were written in Matlab and code-generated for the real-time system. The PPC controlled the gimbal via TCP/IP and interfaced with a home-grown FPGA-based tracker. Image data was collected via a COTS storage system for post processing.
Air Tactical Laser
My primary role for ATL was software lead. The team consisted mostly of self-starters, so my "administration" job was straight forward. Most of my leadership role focused on requirements definition, enforcing consitent code develoment across different subsystems, and verifying interfaces between sub-systems.
My technical role on ATL was the development of the recording sub-system. The sub-system recorded the various video sources and data streams on the aircraft. There were 5 cameras used in the optical train, although none as stressing as the 5Khz ABL camera. I designed a generic engine to handle recording of a single data stream. In some cases, multiple instances of that engine ran on one processor; in others, a processor handled a single camera input. The engine used configuration files to specify the frame rate, image size, and image format of each video sensor.
Lessons Learned
- The adoption of Agile methods in an R&D environment is key. It's what SVS did naturally, but SVS needed some of the data tracking to help with future proposals.
- Requirements reuse is just as abused as code reuse. There are times that code reuse is beneficial (drivers, utility objects, etc), but who would use the same application code for two different directed energy applications?
- Attempting to get a full low-level design of the software is like detailing a schedule. You can't anticipate when, where, or why the slip to the right is going to occur. A flexible (not generic!) architectural design has long staying power.
Fun Facts
- The prototype airborne laser tracker system resided in 3 VME chassis, one of which was a 9-U sized backplane. The flight tracker system now resides in a single cPCI backplane and utilizes just a few FPGA-based processing units.
- On first trip for the C-RAM experiments, SVS tracking system became the gold standard for mortar position.
- When you are out of the middle of nowhere and there are no great features for optical and mechanical alignment, the tops of saguaros work adequately.
Skills
- Embedded Linux/Driver Modules
- Linux tools: Ant SVN Make
- VxWorks
- C
- Java/JNI
- Tcl/Tk
- TCP/IP
- PPC
- Hardware Verification & Validation
- Backplanes: VME PCI cPCI
- Field Test
- MIL Spec Process/Agile Process
Links
Notice: images acquired from various government public-access web sites.