Autonomous Underwater Vehicle Competition

This year, we are participating in the 10th annual International Autonomous Underwater Vehicle (AUV) competition in San Diego, California.

The competition requires teams to design and implement an entirely self-guiding robot capable of carrying out a predetermined underwater mission.

Mission requirements vary slightly from year to year. However, most teams re-entering the competition improve their previous designs and add certain mission specific components as needed.

With the exception of the principle processing unit, all of the other modules contain custom components built since October of last year. This past April we successfully built and tested a mini prototype to accomplish the first part of the mission - the validation gate. That prototype demonstrated the basic functionalities of the AUV - namely navigation, motor control, depth sensor, machine vision and mechanical design. An independent team also designed and proved the successful operation of the passive sonar array.
	AUV Components: + Chassis + Propulsion Control + Power Control + Principle Processing Unit     - Navigation control     - Mission control + Sensors     - Pressure/Depth Sensor     - Temperature Sensor     - Compass     - Inertial Navigation System     - Machine Vision     - Passive Sonar Array + Token Release Module Additional Components + Communication System     - Mid level control     - RS232     - I2C     - USB     - Ethernet + Dockside Computer
Mechanical Our second generation AUV has been largely custom-fabricated by our talented mechanical engineers. The AUV uses twin 22" Lexan cylinders to house the electronics and batteries. Five custom-designed and fabricated brushless thrusters are being used to provide 5 degrees of motion. The top hull will accommodate all electronics, including custom made PCB boards that interconnect through a modular bus backplane system. The bottom hull contains the high power devices, such as the motor controllers, relays, and 2 lead acid batteries. The chassis is primarily aluminium-based, with some non-magnetic stainless steel used for fasteners and the internal rack. The AUV's mechanics are near completion, and a marker dropping mechanism is currently being fabricated, along with the nose and rear cone sections.	AUV Chassis Design
Communication The AUV will incorporate a RS-485 communications network using a Master-Slave architecture. Communication between modules is accomplished by way of sending packets of information from a single board computer (SBC), which is the master, to various sensors and actuators, which act as slaves. The network is half-duplex, poll-based, and uses a customized packet format. The AUV also has provisions for USB connectivity, and some systems will communicate exclusively through USB.	Navigation The AUV uses a C++ based navigation control system running on a Linksys NSLU2 device modified to serve as a secondary single board computer. Navigation control is responsible for the intelligence that autonomously guides the AUV through the competition course based on the information gathered from the various sensors. The modular design of last year's Navigation module allows for the reuse of many lower level commands and tested algorithms to facilitate this year's higher level programming. Added support for the new Doppler Velocity Log, lateral motors and the increased difficulty of the mission objectives are all features projected for this year, expanding on the modular design already in place.
Sonar Acquisition Board	Sonar With the experience acquired from last year, the sonar system is primarily evolving towards more flexibility and reliability. The acquisition board was re-designed to accommodate more diagnostics. In addition, instead of using 8-bits parallel outputs for the ADCs, the new board is based on 16-bits serial transfers between components, reducing the number of wires on the board by a factor of 8, and increasing the precision of conversion by a factor of 2. The basic bearings detection scheme, based on time differences, is still implemented using IIR filters. This year however, a tracking scheme based on Bayesian filtering methods is being developed. Such a design makes it possible to estimate sequentially the position, velocity, and acceleration of an acoustic target, with promising results already observed on simulations. Further testing of the algorithms will be carried out on a Stratix EP1S80 FPGA development kit sponsored by Altera. The onboard FPGA will be a Parallax Stratix SmartPack-EP1S25 board.
Motor Control The Motor Control activates five brushless motors for lateral, depth and forward motion control. These motors were provided by Hacker Brushless, and are more efficient than standard brushed motors. They require a three phase controller with an integrated current control, to limit start-up current draw. The Motor Control uses Proportional-Integral-Differential (PID) controllers to provide closed loop speed control using back-emf (BEMF) for rotor position feedback. The custom designed and fabricated PCBs are already populated, the software is in process.	Power Control The AUV will be powered by two lead acid batteries in parallel, sponsored by GS Batteries, which will be placed in the lower hull. Each custom-made PCB will have its own linear power regulator, creating a modular decentralized power system. A Power Control PCB board will provide power to the SBCs, a hard drive, the cameras, and the FPGA board. The DVL unit requires twelve volts and so will receive fused unregulated power. The power system also incorporates voltage and current sensing circuitry for estimating and displaying to the LCD screen the remaining on-board battery life.
Machine Vision The machine vision algorithms will run on a dedicated TMZ104 500MHz Crusoe single board computer. The algorithms are coded in C++ and are based on the open source image processing library LTI-Lib in order to perform object and colour recognition. The machine vision module will process images gathered by two different cameras: a 3Com HomeConnect CCD camera and a CMUcam2 camera.	CMU Pipe Image
Doppler Velocity Log	DVL For measuring velocity we will be using a Doppler Velocity Log (DVL). This expensive equipment was loaned to us by Sontek. The DVL works by releasing a sound and then measuring sound wave reflections from the surrounding objects, for example, walls or bottom. It uses Doppler's formula to compute the velocity, and outputs it in 3D East-North-Up coordinates. A built-in compass will also provide us with such useful information as roll, pitch, and heading of the submarine.