GPS Deprived UAV Localization

Team Members

• Matthew Manley - [email protected]
• Joseph Krusling - [email protected]
• Tanner Bornemann - [email protected]
• Mahdi Norouzi - [email protected] (Faculty Advisor)

Project Background Description

Unmanned Aerial Vehicles (UAV) often rely heavily on GPS for autonomous navigation. GPS is not available in all flight environments. Indoor autonomous navigation is often extremely limited. There are many methods used to localize a drone indoors and we will explore and compare some of these methods. External Optical localization and WiFi localization will be researched and used to compare with the precision of GPS. External Optical localization will include the use of multiple external cameras used to localize a quadcopter in a known space. WiFi localization will use existing wireless networks to localize a quadcopter in a known space, but has the possibility to work in an unknown space.

Project Problem Statement

GPS deprived environments are notoriously difficult for UAV autonomous navigation. GPS allows the UAV to be localized absolutely. Without GPS the UAV will require conditional localization that is based on what is available in the known environment. We wish to further research the complexities of GPS deprived navigation for UAVs.

Problem Approach

External Optical (EO) and WiFi localization will be explored and compared to known position information. This way any localization data from EO or WiFi can be compared to the already known position information. This will allow for further research into what can be used to improve these methods of localization. Such as a neural network that filters out noise in WiFi localization data or effective use of the scalability of EO with many different UAVs in the same known space.