VisHawk | ECE189 2019-20
VisHawk Project

Project

The VisHawk project is pursuing RF-silent autonomous flight of multirotor UAVs between ships underway at sea. We implement this behaviour by guiding the final approach with Apriltag visual fiducials. Using GPS and a known initial trajectory of the landing vessel, VisHawk closes to within visual range of the vessel. VisHawk then uses the visual marker to produce a safe approach trajectory, updating in real time to consider uncertainties such as sea conditions moving the deck of the landing vessel.

tl;dr? Pursuing RF-silent autonomous flight of multirotor UAVs between ships underway at sea by using computer vision to guide final approach where GPS would be too inaccurate.


Videos

Here's our successful flight demo!

Earlier flight footage and a presentation rehersal are available in our YouTube playlist: https://youtu.be/OROEFxYX0Mw


Team

Matthew headshot

Lead - Matthew Dupree

An okay fellow.

Xihan headshot

Member - Xihan Liu

A great person who's great at drones.

Yingchao headshot

Member - Yingchao Zhu

A great person who's great at drones.


Timeline and Updates

2019 Fall Quarter - Hardware Progress

-- Identified components for and designed COTS part based drone capable of necessary CV
-- Completed testing of all hardware components in isolation, and in communication with their nearest controllers in the block diagram
-- Assembled designed drone

2019 Fall Quarter - Software Progress

-- Identified a visual fiducial system capable of identifying and guiding a drone to moving landing site(s)
-- Identified a communication protocol for the drone flight system
-- Tested a visual fiducial system on target hardware imaging pipeline, with promising results.


More Details

Click here