One of CUAUV’s primary objectives each year is to successfully complete the mission for the Association for Unmanned Vehicle Systems International (AUVSI) Underwater Competition. Teams from around the world compete at this annual event, which takes place each summer at the TRANSDEC Facility in San Diego, California. Completing the competition requires each vehicle to autonomously execute a sequence of tasks while remaining fully submerged. Vehicles are expected to pass through a validation gate, follow colored paths, locate buoys, fire torpedoes, drop markers into appropriate bins, and track an acoustic pinger before surfacing.
For more images of the competition elements through our cameras, visit the Underwater Pictures Image Gallery.
For vehicle perspective videos of completing the mission, visit our Underwater Video page.
For advice for new teams, vist our Advice for new teams section page.
To complete vision tasks, Tachyon has two cameras, one downward and one forward. Depending on where in the mission Tachyon is, it processes video from one or both cameras. All of the machine vision algorithms work in a similar manner in recognizing mission elements. The input image is converted from RGB space into HSV space and split into its three component channels: hue, saturation, and value. Each of these channels is segmented through predetermined thresholds, and the three segmented channels are recombined to form a binary image. Contours are detected in the binary image, and these contours are then run through a set of probabilistic filters and moment analysis to determine the location, orientation, and probability for a specific mission element.
To follow piplines, the computer processes the images by converting them into a binary image and determining the direction of the pipeline using moment analysis. The vehicle then centers itself over the pipeline and snaps to the appropriate heading.
Tachyon docks with the buoy by finding the center of the buoy and adjusting heading and depth until the buoy is in the center of the frame. Once the vehicle hits the first buoy, it backs off, and looks for the second buoy. After hitting the second buoy, it goes over it at the same heading it had before.
Part of this year's mission is driving under "barbed wire," two lengths of PVC next to eachother. Once the vehicle is centered and perpendicular to the pipes and low enough to clear, it drives beneath them while maintaining heading and depth.
Another part of the mission is dropping markers on bins. To do this, the vision software must recognize the shapes of a tank, plane, factory, and battleship. Using moment analysis and probabilistic filters, the vehicle determines what shape it is looking at and whether it is one of its "priority targets." If the shape is a target, the vehicle centers itself over the bin, descends, and drops a marker. If not, it moves on to look at the next bin.
Tachyon first uses the passive acoustic sensor (hydrophone) to determine which of the pingers is active, and selects the correct frequency to track. The hydophone then gives heading and elevation information of this pinger relative to the vehicle. Based on this, the hydrophone guides the vehicle to the recovery zone. More information about the hydrophones is available.