Dive Log Archive
July 21st, 2008
Things are getting very exciting. After lots of work last night, we got two major tasks fully tuned and running. See the videos of the vehicle doing the "Black Jack Table" and pinger below.
July 16th, 2008
It has been over a month since the last entry, but the lack of updates does not denote a lack of progress. Today was the first successful dive of Tortuga 2, which is the full integration of our 2008 AUVSI competition electronics and hardware. The dive operations quickly showed the utility of the new landing gear and the added ease of use the new hardware provides. In addition to that we also got our sonar system to successfully determine the location of full set of pings.
June 17th, 2008
We are pleased to report significant progress on our new Blackfin/FPGA based sonar module. The supporting software is ready, the analog filters work, the Blackfin boots Linux, and the FPGA can read the ADCs. We are looking forward to the rest of the testing and integration. We are all hard at work on all of our systems at biweekly meetings and looking forward to the competition very much!
May 18th, 2008
Tortuga 1.5 is born. A combination of the Tortuga 1's hardware and the electronics stack of Tortuga 2, it represents a large step toward Tortuga 2's completion. In the picture above you can see LEDs from the new electronics glowing through Tortuga's hull. Tortuga 1.5's launch was delayed slightly by some trouble during initial integration of the new electronics with the rest of the vehicle. After repairs and careful testing, Tortuga was back in the water and performing up to par. The new electronics bring many needed safety, monitoring, and maintenance features to Tortuga.
April 26th, 2008
Today we participated in Maryland Day, the University's primary outreach event to prospective students and the entire state. We had a table set up in the Kim Engineering Building (pictured) where we had video of our AUV playing, a simple time vs frequency demo, Tortuga I hardware, and handouts from our new effort with the Autonomous Robotic Speedway Competition. We also set up in the Neutral Buoyancy Research Facility where we demonstrated our AUV in autonomous mode and allowed visitors to drive the robot manually. Both exhibits were highly attended (Approximately 70,000 people come to MD Day every year) and we had a great time interacting with the public and talking about robotics.
April 15th, 2008
Tortuga I has partially completed the pipeline following task from the 2007 course (PDF). Robots have to follow a path marked by a set of long, intermittent orange rectangles, collectively called the pipeline, on the bottom of the arena. Only a fifth of last year's teams were able to follow the pipeline and just 3 teams could track more than one consecutive piece. Tortuga successfully followed two sections of the pipeline. We were not able to attempt three or more today because of dive time constraints. See footage from the pipeline dive below.
This is the culmination of over a year of undergrad member Daniel Hakim's work on our vision code. It achieves its full potential in combination with our 6 degrees of freedom controller, written by graduate student member Joseph Gland. A simple finite state machine guides Tortuga between the pipeline sections.
April 11th, 2008
Today we had a guest lecturer, Shivang Patel, come to speak about image processing techniques and algorithms. He gave a great overview of theory with some practical experience, stories, and video to back it all up. This is one of the first events we have held like this and it was a great success. Thank you Shivang!
April 3rd, 2008
We had our second successful dive with the six degree of freedom Tortuga today. Tortuga 1 has been retrofitted with an additional pair of motor controls and thrusters. This allows us to test the six thruster configuration of the vehicle without having to wait for the other parts of the new vehicle to be finished. This also clears the way to begin testing our pipe following and bin hovering algorithms. Not many teams, including us, were very successful last year at those tasks so we are looking forward to the testing.
In the video above you see the vehicle moving sideways through the water. It is in need of a little gain tuning and thruster calibration, but the performance is currently adequate for our purposes.
March 27th, 2008
Today we have two important things to talk about: a new robotics competition we are participating in, and the competition Tortuga 2's core electronics. The Robotics @ Maryland club is holding a kickoff meeting this Thursday (March 27, 2008) in AV Williams Room 2328 at 6:30 PM. The meeting is for a new autonomous robot competition, the IEEE-RAS-DC, to be held in the DC area this summer. The competition requires teams of students to design, build, and program small autonomous vehicles to compete an outdoor course as fast as possible. If you are interested in robotics and you would like to get involved in a very active club, then this is a perfect opportunity. Please RSVP at:
all are welcome. If you wish to learn more, visit the site linked above or e-mail the competition organizer, Melanie Vida, at 
.
Along with that exciting opportunity to start on a new project, we have made progress on Tortuga 2's electronics. After working into the early morning Nick Limparis and Steve Moskovchenko finished and integrated the new balancer board with the other three new boards and the backplane. This is a culmination of months of work by the electronics team, and is large step toward getting Tortuga 2 into the water.
March 4th, 2008
Tonight Tortuga participated in its first research project. Aerospace engineering student Madeline Kirk had us outfit Tortuga with a docking apparatus that she is designing for her honors research project.
Madeline is an undergraduate who works at Space Systems Laboratory (SSL). The docking device is designed to be used on SCAMP, a six-degree-of-freedom robot operated by SSL.
The first docking attempt had some hiccups. We accidentally plugged the vehicle's Ethernet tether into a 10 Mbit network switch instead of our usual 100 Mbit switch. This decreased the responsiveness of the joystick control and the quality of the video from onboard cameras. Next, the right thruster cut out. Fearing a blown motor, we hoisted Tortuga onto the deck and inspected the misbehaving thruster. We found strands of fine fiber wound about the impeller shaft, which we think came off of our webbing straps.
After several hours of meticulously disassembling, cleaning, and reassembling all four thrusters, we started our second attempt. With a few tweaks to the joystick controls, we were able to make several successful docking runs. The docking device has a unique self-aligning design that allows the vehicle to mate with the platform even when approaching from slightly oblique angles. You can see this in the video below:
February 26, 2008
Steady progress in sonar and vision continued today. Our hydrophones picked up signals from our pinger, which is so loud that we can hear it anywhere on deck even though it is submerged several feet below the water's surface. We need to tone it down to match the intensity level from the competition's spec.
Our new hydrophones pick up clear signals in competition's 22 kHz-30 kHz pinger frequency range, though today we generated and detected ultrasound pulses up to 240 kHz and beyond. We also observed surface echoes, wall echoes, and acoustic interference patterns (probably from the Lloyd's mirror effect).
Our AUV, Tortuga I, has fully completed the red light task from last year's competition. Only a quarter of the teams completed this objective in 2007 and just 3 teams in 2006.
Above is a video from the vehicle's forward camera, played back at several times the original speed. The vehicle starts out by searching for the red light, finds it, makes contact, then surfaces to end the mission. The small blue square marks where the vehicle "thinks" the red light is.
February 21, 2008
More breakthroughs! We strung a rope across the tank and hung our blinking red light from it. By pulling the rope back and forth, we could move the red light horizontally through the water at any speed. Tortuga tracked the red light continuously by adjusting its azimuth (compass angle) and forward speed.
Also, we heard the first sounds from our pinger. We sent burst sine waves from our signal generator through a COTS power amplifier to drive the pinger. Within less than an hour of trial and error we started hearing strong, clear pings throughout the tank. We also were able to play DC101 underwater.
February 19, 2008
Today we acquired the first signal from our new hydrophones.
February 12, 2008
After a few weeks of video collection, vision testing, and software integration we have the vehicle autonomously tracking the red light! The vision algorithms developed by Daniel Hakim fed the relative heading to red light into the motion system which turned the vehicle toward it. Integration was sped up by the use of the simulator confirm the correctness of the tracking algorithms given an ideal vision system.
The dive was eventful. We encountered swapped thrusters cables, network errors, and a blown chip by the time it was finished. By next meeting, we will replace a blown logic chip on the power board and add some shielding to our network cable to eliminate some connectivity issues.
When all was said and done we had about 30 minutes of light tracking fun.
January 21, 2008
Tortuga 's AI guided the vehicle through the starting gate, allowing us to check off a number of tickets in the version control system. We logged data to tweak the depth controller. We also tested a preliminary version of the vehicle GUI and found that it was mostly functional. We can now monitor desired direction, depth, roll, pitch, and yaw, all in a compact package. Kudos Jon Speiser and Joe Lisee for the work on the GUI.
The "south bug", the condition where the sub would start oscillating ~5 degrees when pointing south has been fixed using the simulator as a testbed. Finally we started encountering some errors that we had never experienced before and after a bit of trouble shooting determined that the error were being caused by low voltage on the battery cluster. Tickets have been made to put in to the ticketing system to start using the voltage monitoring hardware in software to alert for low power conditions. Finally the moment you all have been reading for. Pictures!!!
Septempber 4, 2007
The team has been meeting bi weekly to setup a demo for our sponsors and for the University's annual First Look Fair where all student groups set up information tables and demos on the grass mall. We are planning our new member kickoff for sometime near the end of September.
