Flying Machine Arena

We learn best through direct experience because there are real limits to our ability to process complex instructions.
About the Project
Research Objectives
Documentation

Thesis & Semester Projects

Contact

Research Objectives Adaptation and Learning
How can we develop the algorithms that will allow a vehicle to learn from experience, and yet not damage itself in the process? Avoiding collisions, despite disturbances in lighting conditions and communications, is crucial with multiple vehicles flying in a single airspace.

Our aim, however, is to go beyond collision-avoidance. We hope, over time, to see our system guiding a dozen or more vehicles into complex, acrobatic flight formations.

Camera System
The camera system must be able to track a dozen or more quadricopters over a 10 X 10 X 10 meter airspace.

Vehicle Requirements
The vehicles must be agile, reliable and they must be able to fly autonomously. They must also be robust enough to withstand crashes and collisions, and yet still be safe for human beings (i.e.: lightweight, flexible propellers).

Automatic Platform
The goal is for the Flying Machine Arena to be a fully automated platform, where camera calibration, flight control and vehicle charging are all regulated by the system.