The Magazine of IEEE-Eta Kappa Nu October 2017 - 23

A Student-Built Fixed-Wing UAS for Simulated Search-and-Rescue Missions

I. Introduction
In today's world, the word drone means a variety of
things to different people. To the general public, a drone
is simply a nifty gadget or toy that is the perfect gift for
any hobbyist or middle-school student for Christmas.
To some, it is viewed as a necessary tool for capturing
breathtaking views of landscapes, city skylines, and events
for great video productions. However, we at University
of Hawaii Drone Technologies (UHDT) view drones,
or unmanned aerial systems (UAS) as they are more
appropriately labeled, as effective tools for solving realworld problems [1]. The real-world problem that UHDT is
specifically addressing is search and rescue (SAR).
Current solutions for SAR missions require costly
manned aerial vehicles be deployed to the area of
interest. A SAR vehicle deployed to rescue an individual
costs the US National Park Service up to $220,000 [2].
Not only is the high operating cost a factor, but risks
involving SAR crew safety is a concern as well. High risks
to safety of both crew and target could compromise the
SAR mission. Reducing operational costs and safety risks
for SAR missions will lead to a more effective solution,
motivating the need for a SAR-specific UAS [3]-[5].
Due to the unmanned aspect of a UAS, the safety and
operational costs for a rescue crew are no longer ethical
or financial concerns in SAR. Removing the onboard
human operator reduces the required lift; therefore, less
fuel is consumed and less powerful, and less expensive
propulsion methods can be considered.
For SAR, the risks and consequences are real. Using
real SAR missions as a testing method would be
inappropriate. Finding an appropriate means of testing
a SAR UAS during development is a must [6]-[8]. The
Association for Unmanned Vehicle Systems International
(AUVSI) Student Unmanned Aerial System (SUAS)
competition simulates a SAR mission complete with
waypoint navigation, target search and identification,
and air delivery of a payload to a stranded individual
[9]. The annual competition provides a comprehensive
simulated SAR mission without the risks involved with an
actual SAR mission.
This project was carried out as part of the University
of Hawaii's Vertically Integrated Projects (VIP) Program
[10]. The goal of the VIP Program is for students to work
on a project over several years, so that a student rises
in experience to the senior level and passes down the
knowledge to new students joining the project. The team
was comprised of approximately 20 undergraduate
THE BRIDGE // Issue 3 2017

students in computer, electrical, and mechanical
engineering of various academic standings ranging from
freshmen to seniors.

II. System Overview
The system can identify specific targets of interests, such
as alphanumeric targets of various shapes and colors
in a specified area, and safely deploy a package to a
specified target. To perform these tasks, the system is
comprised of the following subsystems: air frame, air
delivery, flight controller, image capturing and processing,
communications, and ground control station (GCS). Figure
1 depicts the system-level diagram of the UAS. There are
two major components: the aircraft and the GCS. The UAS
must be able to have the appropriate usable flight weight
and internal volume to accommodate all of the electronic
components needed to complete the tasks of autonomous
flight, target classification and localization, and air delivery.
For the simulated SAR mission, the UAS first navigates
a set of waypoints provided by the competition. The
UAS then searches a predefined 370,000-m2 area for
targets by capturing images every four seconds while
flying a linear, cross-grain search pattern at an altitude
of 70 m. While taking pictures, the UAS processes
the images in real time on the onboard Raspberry Pi
computer. The images are also sent to and processed
on the GCS during post processing for redundancy. The
processing algorithm filters images with possible targets
and further analyzes them for color and shape of the
detected target. The algorithm identifies at least two of
the following characteristics: shape, background color,
alphanumeric symbol, alphanumeric color, and GPS
coordinates. Next, the UAS leaves the search area to fly
to the specified air-delivery location where it deploys an
8-oz water bottle payload. Lastly, the UAS completes

Fig. 1. System-level diagram of UAS
23

Table of Contents for the Digital Edition of The Magazine of IEEE-Eta Kappa Nu October 2017