A.L.V.I.N.N.

(Autonomous Learning Vehicle Integrating Neural Networks).

Project Statement

This project is in collaboration with Rockwell Collins to use computer vision, machine learning, and neural networks to have a drone detect an airport runway or objects in the air. The project has two teams assigned to it and our team will focus on detecting objects in the air. Classifications of the objects that could be detected in the air include airplanes, helicopters, drones, birds, and stationary items such as buildings. This will require the team to learn about computer vision, machine learning, and neural networks to identify and implement the software and hardware requirements capable of performing these tasks. Furthermore, part of the project will consist of determining which machine learning algorithms are particularly suitable for our object recognition task. Computer vision tools will be used to process images from a picture or video stream to detect key features of objects. Machine learning and neural networks will then be incorporated to perform teaching operations for identifying different objects or distinguishing between similar objects based on those key features. Since this problem is interdisciplinary in nature, we will also have to investigate to what extent traditional computer vision techniques could be incorporated into our solution.

Purpose

Rockwell Collins is an aviation electronics company dealing with military and commercial aircraft.  Rockwell Collins could use this technology to introduce new products that could have multiple uses for their customers.  Although we are starting simple with basic object detection this could develop into technology that the military could use to survey an area before sending troops in, detect enemy forces, identify bombing locations, or drop locations for aid packages.  Commercial uses could be finding hotspots in forest fires, locating remote wreckage sites, finding lost hikers, or finding survivors of some disaster. The diversity of these different use-cases shows the extensibility, power, and usefulness of our project.