top of page

Phoenix

3rd at 2019

RoboSub

General Overview

The team’s autonomous underwater vehicle, named Phoenix, includes 10 thrusters as well as a passive sonar, Doppler velocity log, fiber optic gyroscope, and two cameras. Phoenix is capable of precise autonomous navigation, manipulating objects, locating the position of an acoustic signal, and classification via vision processing. At the 2019 RoboSub Competition, Phoenix placed 3rd internationally and first in the U. S.

Phoenix.gif
Follow Us to
Stay
Updated!
  • Instagram
  • Facebook
  • Twitter
  • LinkedIn

Mechanical

Phoenix-Main-Hull

Hull Design

Phoenix integrates four hulls; the main hull, two battery hulls, and a seperate hull for the DVL (Doppler Velocity Logger). They are made using clear acrylic tubes for visual LED feedback during the coding and testing phases.

Battery-Hull

Internal Lattice for the Main Hull

The frame was built to connect the Main hull and be powered by one battery hull. All electrical wiring and components are housed within the main hull. 

Phoenix-Frame
Mechanical

Electrical

Electrical

Customized Boards:

AUV-Main_edited.jpg

Mainboard

The Mainboard communicates by sending sensor information to the Jetson using C++ code. The custom PCB allows the Jetson TX2 to handle higher-level tasks including mission management, control loops, network communications, vision processing, and data logging. This PCB utilizes an Arduino Pro Mini and Fiber Optic Gyroscope (FOG) to support navigation, positioning, and maintaining the AUV's heading.

Power Supply Board

The AUV uses 2 LiPo batteries to separately power the computer and thrusters. With having separate power supplies, the team can test the thrusters and electronic components separately. Additional code was used to design a power down sequence to longevitize the components on the AUV using a microcontroller and MOSFET switches. 

20190621_194759 (1)_edited.jpg

Software Testing

Screenshot 2024-06-27 190949.png

Darknet Yolo

The team uses Vision to allow the AUV to detect and classify objects underwater using the MIPI Camera and Darknet Yolo. Darknet was used due to the 120% increase of accuracy in object detection when compared to YOLO V5. The AUV reads data from Yolo and makes a decision based on the distance, type of object, and accuracy of object detection. 

20190614_183449.jpg

Simulation Testing

SOLIDWORKS CAD Modeling

The team used SOLIDWORKS' CAD Modeling software to simulate and build components on the AUV. SOLIDWORKS helps to simulate the AUV's center of gravity, buoyancy, thrust, and drag. Specifically, the CAD Modeling simulated the unbuilt AUV's thrusters and components. 

20181022_172147.jpg
Software
Testing

Misc. Testing

Acoustics Testing

The AUV uses a Subsonus hydrophone array  to position itself underwater.  The AUV positions itself based on a pinger placed in the water. The team placed the pinger in different locations to test the accuracy of the AUV.

Buoyancy Testing

Buoyancy Testing

The AUV is placed into the water and should be neutral in all directions. This would affect the movement of the AUV as it travels underwater. Testing ensures that the AUV is neither positive or negative. Testing revealed the built AUV was positive and 3D Printed Shot Holders were added to bring the AUV closer to neutral. 

bottom of page