The University of Texas at Dallas fielded a small team for the 7th International Autonomous Underwater Vehicle Competition. Despite it being their first time, the UTD team did quite well, scoring 9th among 18 teams, beating many second generation robots and even outscoring teams from schools with highly respected robotics programs like USC. For more photos and video of the event see the UTD team's gallery or the Cornell Team's gallery. Links to all of the participating teams can be found on the AUVSI site. Ed Paradis of the UTD team talked to many of the other first time competitors during the event and collected some interesting data on the different approaches taken to the problem. Read on for details.

Ed's survey of teams

During the competition, I talked to a few of the other first year teams who were competing. The atmosphere was very open, and so I was able to ask each team what they were using in their design. I asked each team what processor they were using, and also what sensors they had connected.

Many of these teams, including my own, had a very short development time. I found it interesting to see how each team attacked the problem. Systems varied from single microcontrollers to dual processor Pentium 3 cards.

Brigham Young University, Hawaii

• BIC – Savage Innovations
  • Visual Basic
  • easy interface
  • random glitches – leave it off and it will work
• compass
• camera – cmucam2
• depth sensor – didn't arrive in time
• Anuj Sehgal, Jason Kadarusman, Christa Kadarusman
• $500, 2-3 months

San Diego State University

• Daniel McKellar
• FPGA, spartan, Pegasus Development Boards
• custom boards for tilt, depth, and motor controller
• software: Xilinx, free version
• graduate student and adviser would cut boards
• each person would focus on a system
• not much experience
• learning curve
• fiberglass, carbon fiber, Kevlar
  • foam form, make shell around
  • eat out foam with acetone
• purchased at “Diversified Materials”

AV High school

• Carl Jantzen

University of Southern California

• Dual Pentium 3 “Rocky 3742EV”
• 30gig HD
• three cameras, one movable
• Biologically inspired visual system
  • how humans and primates view surroundings
  • finding “salient points”
  • ilab.usc.edu – open source vision library
• 21 year old power section, modified ROV
  • completely rewired to control motor

Southern Polytechnic State University

• distributed uC
  • phytec C591, 8bit, 8051
  • donated
• Chare Sanders
• two CMUcams
• two SQ03 hydrophones – Sensortech
• pressure sensor -omega PX26
• unimplemented Futaba Gyro, 2 axis
• active ballast
  • pump, limited amount, hard to seal
• 8 weeks development time
• 30lb Mercury trolling motors
• two main motors, active ballast for vertical movement
• lead acid, 9Ah, 12v
• class offered as upper level elective
• no programmer, but many people with job experience – programming,welding, diverse group
• 15 to 16 people, 2 people per subsystem
• custom ring network rs232 4800 baud
• no time for integration before the competition
• next time:
  • smaller
  • start earlier
  • more sponsors

University of Ottawa

• www.site.uottawa.ca/arise/
• interested in rotomotion IMUs
• Siraj Sabihiddin
• SBC Transmeta Crusoe 500MHz
  • Tri-M B.C. (PC 104)
  • TMZ104
• 144megs RAM, 256M DOC, 1.6g Laptop HD
  • DOC – nothing but trouble
  • good board: low power, good peripherals
  • another USB would be nice
• network card: PC104
• PC104 VESA card
• tethered connection – Ethernet
• pressure sensor: MSI MSP600 (MSD600?)
• digital compass: Honeywell HMR3000
• two cameras
  • 3 com connect (support auto gain, which they use)
  • very good low light response – CCD
  • discontinued, but good investment
  • logitech quickcam 4000: bad support
• hydrophones: donated by DRDC (Canadian Military)
  • measuring time of arrival
  • FPGA, digital filter, and data capture card, Altera Stratix Platform DSP

• modified bilge pumps
• unused IMU
• waterproof – 'impulse' donated by Romor
• battery: lead acid, 18Ah, 12v
• would redesign hull
• hull is made of ABS tubing
  • custom motor controllers

University of Central Florida

• 100 MHz Prometheus PC104
• P4 2.4 Ghz ' Cappuccino Mocha'
• photodiode with lens to track light
• Aquaview hydrophones, no conditioning circuit
• Honeywell HMR-3000 for compass
• ONAVI gyropack
• P4 used for vision system, Matlab control software
• custom humidity sensor: cut capacitor
• top and bottom props counterpitched
  • yaw achieved by difference between props
• Matlab and Simulink good for rapid prototyping
• Next year:
  • same propulsion
  • different motors
  • conditioned power system
• Fischer waterproof connectors
  • long lead time
  • fulyton connectors (?)