Date

2016

Document Type

Thesis

Degree

Master of Science

Department

Mechanical Engineering

First Adviser

Hart, Terry

Abstract

With the advancement of single UAV flight control there is a clear understanding of the importance for future group UAV distributed control. This will in turn lead to scenarios of “smart” communication between UAV teams. This hierarchal chain reaction type control of UAV’s will provide more enhanced real time flight pattern optimizations without the slow interactions of a UAV to a computer (aka a human). By relaying to just the “Master” from the ground station to switch trajectories the human interaction never needs to go any further to update the flight formations of the rest of the group members. This type of “swarming” with UAV’s is only possible with the correct hardware and software improvements. This is especially true when the trend for UAV “groups” are to be higher in number and therefore much smaller (the size of an iPhone). Apart from the already existing ways to manipulate flight paths – hardware including better GPS locating and new censoring technologies for collision/spatial recognition – software limitations are apparent to be the next large hurdle. To accomplish such interactions between UAV’s, optimal flight patterns must be attained before any inter-communication can be implemented. Current designs follow the traditional PD/PID control schematics, but these lack the requirements to correct for real-world disturbances. Kalman filtering control design on the existing architectures of the Ardupilot and Lisa/S flight controllers was implemented to produce the most accurate flight paths of the UAV’s.

Share

COinS