System Architecture Study Global Hawk Unamanned Aerial System (UAS)
HighSpeed_Stealthy_PayloadFocused_VTOL_UAV
1. Designing a High Speed, Stealthy, and Payload-
Focused VTOL UAV
Michael Becker and David Sheffler
University of Virginia, mcb9ze, das2jt@virginia.edu
Abstract - Traditional intelligence, surveillance, and
reconnaissance (ISR) missions conducted by the military
and intelligence community typically are associated with
high costs, slow response times, and inflexible designs.
Current UAV technology has the potential for use in
ISR, but typically copters are limited in speed,
endurance, and payload capacity. As the technology
develops, there is a growing interest in developing small-
scale UAVs with vertical take-off and landing (VTOL)
capability. A VTOL-capable UAV with autonomous and
high-speed payload delivery will prove useful in ISR
missions. The goal of this project is to develop a proof of
concept design for a fixed-wing UAV with these
capabilities along with improved endurance and stealth.
An additive manufactured airframe and commercial off-
the-shelf (COTS) components allow for low cost and
rapid re-configuration. The completed design will
incorporate multiple servos for a variety of control
surfaces, a rear tilting motor for forward and vertical
flight, an embedded lift fan system for vertical flight
stability, and an Android-based control system for
autonomous flight. Within this cross-disciplinary, multi-
year project, the objectives for the 2015-16 year are: to
begin understanding the flight controls and algorithms
required for VTOL flight, to design the aerodynamic
shape and final airframe, to understand how to integrate
the software and hardware with the airframe, and to
develop a flight rig to test VTOL transition capabilities
using RC. Results presented in this paper include details
behind the design and layout of electronic components
and the building and testing of the flight rig. Hardware
decisions for this aircraft were based on the following
design requirements: control system allowing for both
fixed wing and tri-copter airframes, ability to transition
between vertical and horizontal flight within three feet
after vertical lift off, sufficient battery power to allow for
reasonable flight endurance, ease of packaging within the
fixed wing airframe, compatibility with servo
requirements, and a total system cost of less than $5K.
The team found suitable COTS components that satisfied
the program objectives at low cost. Functional testing of
vertical and horizontal flight transitions with the rig is
currently underway and autonomous flight capabilities
are under development.
Index Terms - Additive Manufacturing, Android, COTS,
UAV, VTOL
PROBLEM BACKGROUND AND IMPORTANCE
UAVs are becoming an important asset in military ISR
operations. According to the Department of Defense’s
(DoD) “Unmanned Systems Integrated Roadmap”, federal
agencies plan to spend $14.5 billion on unmanned systems
over the next three years, of which $13.1 billion will be
allocated to UAVs [1]. In recent years the military has been
looking towards UAVs to provide an undetected and
efficient method for delivering small quantities to soldiers
out in the field [2]. Current technologies like copters lack the
stealth, speed, and endurance to fly undetected during
missions. Fixed-wing configurations like the RQ-11 Raven
provide stealth and endurance for ISR missions, but lack the
capability for efficient payload delivery [3]. For this reason
small-scale UAVs with VTOL capability are gaining more
attention.
Current military ISR missions are characterized by high
costs, slow response times, and inflexible designs. Payloads
and sensor technologies are evolving much faster than
vehicle platforms causing an increase in cost and time spent
for replacement and maintenance. The DoD is interested in
designing airframes with modularity in order to support the
rapid re-configuration of components and airframes [1]. One
possible solution is to use additive manufacturing for ease of
replacement and re-design on site. New airframes and
components can be 3D-printed in a matter of hours allowing
for quicker turnaround times. Commercial off-the-shelf
(COTS) components will prove useful because they are
available at low costs. This project involves designing and
building a fixed-wing VTOL UAV that demonstrates these
advantages.
PREVIOUS WORK
The MITRE Corporation (MITRE) previously worked on a
similar project called the Razor Program with the University
of Virginia in 2012-2013. This program developed a 3D-
printed flying-wing UAV platform with a 3D Robotics
(3DR) Pixhawk autopilot system and an LG Nexus 5
Android device that utilized MITRE’s Android Control and
Sensor System (ACSS). The aircraft cruised for 45 minutes
at 40 mph and had a total fielded cost of $2 K. The ACSS
allowed for simple automated flight via pre-programmed
missions. By using a reconfigurable platform, the Razor
employed modular, snap-fit wing and fuselage components
[4]. However, the aircraft was designed to be hand-launched
and did not have the ability to conduct VTOL transitions or
2. deliver payloads. This project has grown into a cross-
disciplinary, multi-year program (2015-2017) after receiving
funding from a federal agency. The ultimate goal for the new
program is to prototype a 3D-printed flying-wing UAV with
additional VTOL capability and autonomous payload
delivery. A simultaneous localization and mapping (SLAM)
algorithm allows for autonomous landing site detection and
situational awareness. Refinements to the aircraft include:
improved speed and stealth, a new autopilot system with
VTOL transition programmability, two embedded lift fans, a
rear tilting motor, and clam-shell elevons.
Prior to 2015, motor designs were completed and trade
studies were conducted in determining the optimal batteries
and ESCs for the system. After characterizing the motors
and measuring power data, it was found that two 120mm
fans would provide up to 4.8lbs of thrust at 40A each for
hovering flight, and a 70mm rear motor would provide
5.3lbs of thrust at 70A for hovering and horizontal flight. In
order to achieve desired flight capabilities, it was determined
the aircraft weight should be no heavier than 12lbs with a
wingspan of 60 inches.
GOALS AND OBJECTIVES
The goals of this project for 2015-2016 were as follows:
1. To complete final airfoil and airframe design for
the UAV using XFRL5 aerodynamic simulation
and SolidWorks
2. To begin implementing a custom VTOL airframe
for the Pixhawk autopilot system using PX4 open
source flight stack software
3. To identify electronics and hardware required for
the UAV and to outline their integration with the
airframe
4. To develop a transitional-flight rig integrated with
motors and electronics to test flight capabilities
using remote control (RC) and a semi-automated
VTOL transition
This paper will focus on Goals 3-4 as other members of
the team were responsible for Goals 1-2. Objectives were
formulated for these two goals:
Electronics Objectives:
1. Ease of assembly within airframe with an even
weight distribution
2. Sufficient battery power to supply components
and achieve reasonable flight endurance
3. Safe and effective distribution of power
4. Use of low-cost COTS components
Transition Rig Objectives:
1. Functional testing of vertical and horizontal
flight transitions within three feet off ground
2. Lightweight frame with components mounted
in intended configuration
3. Use of simpler autopilot system for
understanding motor behavior for VTOL
These objectives guided the design and configuration of
the electronics and transition rig. The ultimate goals for the
program also guided decision-making.
PROJECT DETAILS
A CAD model of the UAV VTOL system is shown in Figure
1. This paper will cover the electronics design and transition
rig development and testing portions of this UAV program.
FIGURE 1
SOLIDWORKS MODEL OF UAV SYSTEM
I. Electronic Components
The new UAV is larger than the Razor, but there are
still space constraints that were considered. The 3DR
Pixhawk and accompanying components (radio, receiver and
transmitter, GPS, etc.) used on the Razor are also being used
on the new configuration. Instead of an LG Nexus 5 Android
phone, the new configuration will use a Samsung Galaxy S6
device. A IOIO board is to be used as an input/output board
for the Android device. These components are to be
assembled in a 3D-printed component tray mounted on the
front of the aircraft via snap-fits. In selecting the remaining
COTS components (batteries, ESCs, power distribution
board, servos), alternatives were primarily evaluated on their
size and cost. Becoming aware of what the open-source RC
community was doing assisted in selecting dependable parts.
Figure 2 shows the general architecture and information
flow between components.
FIGURE 2
ARCHITECTURE AND INFORMATION FLOW
3. The Pixhawk autopilot will maintain control of the
UAV’s stability. The RC transmitter will sends signals to the
receiver onboard the aircraft. A ground control station
(GCS) software application will display real-time data about
the UAV’s position through wireless telemetry and allow the
user to create pre-programmed missions. The Android phone
is to serve as an embedded control system utilizing MITRE’s
ACSS to allow for autonomous flight capabilities. The IOIO
board will allow data signals to be sent between the Android
device and the Pixhawk.
Power Distribution: After characterization of the
propulsion and lift fan motors, the lift fans were found
to require a maximum of 40A and the rear motor 70A.
60A and 110A electronic speed controllers (ESC)
respectively were chosen to regulate the voltage (and
speed) to the motors. It was determined the max
amperage required by the system was around 200A. A
6S, 22.2V, 55C, 5000mAh battery has a 275A discharge
rate. Typically it is desired that the system’s discharge
rate be twice that of which is required in order to extend
the battery’s life. This would also achieve greater flight
endurance. Therefore the new configuration would
include two 6S batteries in parallel, which would supply
a constant 22.2V and have a 550A discharge rate. A
power distribution board (PDB) from Gryphon
Dynamics was chosen to distribute the high current
among the motors and supply 5V to the Pixhawk using
an onboard BEC. Figure 3 shows the power distribution
among the different components. 12-14AWG wire and
Dean’s/XT90 connectors were used in connecting
components.
FIGURE 3
POWER CONNECTIONS
Airframe: The configuration of the electronics and
motors in the airframe is shown in Figure 4. The
Pixhawk, Android phone, and IOIO will be neatly
mounted in a 3D-printed “component tray” that slides
into the nose of the aircraft. The three motors will be
mounted in a tricopter configuration, with the batteries,
ESCs, and distribution board assembled in between.
Nine Futaba servos are to be used for the final
configuration: four for clamshell elevons (roll and pitch
stability), three for rear motor tilt about two axes of
rotation (yaw stability and VTOL transition), and two
for cover doors that shield the lift fans during forward
flight.
FIGURE 4
ELECTRONICS CONFIGURATION
Control System: The Pixhawk autopilot developed
through the PX4 open hardware project uses advanced
processor and sensor technology for the control of
autonomous vehicles. Onboard the UAV a 915MHz
radio will be used for wireless communication with the
GCS, a GPS/compass will provide the autopilot with
positioning data, and a built-in accelerometer,
gyroscope, and magnetometer will provide aircraft
stability. A Spektrum receiver will also connected to the
Pixhawk through a pulse position modulation (PPM)
encoder, which allows for RC control via a Spektrum
Dx7s transmitter. GCS platforms provide airframe
options for users depending on the aircraft they are
flying. For this project the team is accessing the PX4
software environment and writing unique code, mixer
methods, and PID gains to create a customized airframe
that will include VTOL capability. Current interest in
VTOL flight among the PX4 community is providing
the team with the Firefly6 airframe [5], a tilt rotor
VTOL aircraft, which is being used as a reference as
they implement the new UAV configuration.
VTOL Transition: For an outbound transition, all three
fans will be activated with the rear motor tilted
downwards. The UAV will hover just like a tricopter
with stability about pitch and roll controlled by
adjusting the voltage sent to the motors. Yaw stability
will be achieved by adding a second degree of rotation
on the tilting motor. In order to transition to horizontal
flight, the lift fans will gradually decrease in power as
the rear motor is tilted backward. Clamshell elevons are
to provide roll and pitch stability during forward flight.
Cover doors controlled by servos will shield the lift fans
during horizontal flight for improved aircraft
4. aerodynamics and speed. For an inbound transition, the
aircraft will slow to a cruising speed with the assistance
of extended elevons which act like flaps. The order of
operations is essentially reversed until the UAV returns
to hovering and can land safely.
II. Transition Rig
The same electronics setup just described was used for the
transition rig, but the Pixhawk was replaced with a simpler
autopilot system. The primary goal of the rig was to test
basic VTOL transitions using the two lift fans and tilting
rear motor. Control surfaces and a second degree of motor
tilt would be integrated on the rig at a later time for more in-
depth future testing. Figure 5 shows a photo of the
completed rig. In order to achieve a lightweight frame,
modeling foam was the primary material used in addition to
plywood and a 3D-printed mount for the rear motor.
FIGURE 5
TRANSITION RIG
Airframe: The airframe was cut out of 2-inch thick
modeling foam. The motors were securely fastened onto
the foam model using plywood and screws. A 3D-
printed mount was used to provide rear motor tilt from
two servos. The motors were mounted so as to place the
COG one unit of measurement away from the lift fans
and two units away from the rear motor. When the
electronics were mounted, two metal weights were
applied to the back of the rig to account for wing
masses. After characterization of the motors the lift fans
were found to provide a maximum thrust of 4.8lbs each
and the rear motor 5.3lbs. A free body diagram (FBD)
shown in Figure 6 was used to determine what hovering
thrust was required for each fan and to confirm that
there was no moment on the aircraft. The weight of the
rig was measured on a scale to be 10.86lbs. Calculations
confirmed that the amount of thrust required for each
fan was 3.6lbs, which is 75% of the maximum thrust
capable for an individual lift fan and 68% of the
maximum thrust capable for the rear motor. This was
deemed sufficient for our design because the aircraft
will primarily be flying in forward flight.
FIGURE 6
RIG FBD FOR HOVERING
Control System: The autopilot system chosen for the
rig was HobbyKing’s KK2.1.5 multirotor flight control
board. An open-source firmware called
OpenAeroVTOL (OAV) [6] was uploaded to the board
which allowed for full transitional control and mixers
without any programming required. Various parameters,
settings, and PID gain values are entered into the system
through a menu driven interface on an LCD screen. The
flight controller includes an onboard accelerometer and
gyroscope for stability control. The open-source VTOL
firmware, low price, and small size of the KK2 ($20)
made this flight controller the best option for the
transition rig.
VTOL Transition: The flexibility and customizability
of the OAV firmware allows for the transition between
fixed wing and tricopter configurations. The firmware is
based around two modes, a hovering mode (P1) and a
forward flight mode (P2) [6]. Using OAV, a customized
transition sequence between these two modes was
designed with the flip of an AUX switch on the
transmitter. Output mixers were used to adjust the
amount of throttle for each motors and pitch (rear motor
tilt) that occurred between the transitions. Throttle
curves, managing the P1 and P2 throttle volumes, were
selected to be sine curves so as to ease the VTOL
transition.
CURRENT STATE AND NEXT STEPS
The electronics objectives for safe distribution of power and
the design of an organized configuration were achieved.
Simple calculations show that using two 6S, 22.2V batteries
should provide sufficient power for all components while
allowing for reasonable flight endurance, but further flight
testing will verify this. The transitional rig was successfully
built and the KK2 has proven to be a suitable flight
controller for transition testing as the Pixhawk airframe
development continues. Flight testing has proved the
compatibility of the electronics and motors, and it confirmed
that the wired/soldered connections and the power
distribution board were operating correctly. This was an
5. important step because it allowed the team to troubleshoot
any issues and find any gaps in the complex array of
electronics. By verifying the electronics assembly was
operable, the future integration of the components with the
Pixhawk and mounting within the additive manufactured
airframe will be completed efficiently. Functional testing of
vertical and horizontal flight transitions with the rig is
currently underway as we attempt to perfect the PID gain
parameters to create a smooth transition.
The next step for this program is to add controlled
surfaces to the transition rig to more closely emulate the
final aircraft and improve flight testing performance. When
sufficient flight testing has been completed and the team has
a good understanding of what is required to complete VTOL
transitions, the electronics will be flight tested for longer
periods at a time in a 3D-printed airframe. When the
customized Pixhawk airframe is completed the Pixhawk will
replace the KK2 and automated flight will be introduced.
Further development of the VTOL UAV will
demonstrate possibilities for military operations. The
capability to safely deliver payloads to troops in the field has
the potential to save soldiers’ lives. Other applications may
include more extensive and stealthy surveillance and
reconnaissance missions. The concept of an additive
manufactured airframe allows for reconfiguration and
redesign out in the field, eliminating high costs and slow
turn-around times. As 3D-printing technologies further
develop and become more affordable, they will become a
valuable asset in conducting military operations.
ACKNOWLEDGMENT
We would like to acknowledge the other University of
Virginia students involved in this VTOL UAV program who
were primarily responsible for Goals 1-2 as listed in the
“Goals and Objectives” section: Ryan Porter, Andrew
Freschi, Paul Hughes, Leandra Irvine (Department of
Mechanical and Aerospace Engineering), Daniel Coo and
Adrian Gloria (Department of Computer Science).
REFERENCES
[1] Department of Defense (2013), “Unmanned Integrated Systems
Roadmap.” Retrieved from http://www.defense.gov
[2] Glade, D. (2000). “Unmanned Aerial Vehicles: Implications for
Military Operations.” Center for Strategy and Technology Air War
College. Retrieved from http://www.au.af.mil/au
[3] AeroVironment, Inc. (2016). “UAS: RQ-11B Raven”. Retrieved from
http://www.avinc.com
[4] Balazs., M., & Rotner, J. (2013). “Open, Commercial Technologies
Lead to Cost-Effective Reconnaissance Solutions.” Retrieved from
http://www.mitre.org
[5] PX4 Autopilot. (2016). Retrieved from https://pixhawk.org
[6] Thompson, D. (2015). OpenAero-VTOL Firmware for KK2.1x
Boards (V1.3). Retrieved from http://www.rcgroups.com
AUTHOR INFORMATION
Michael Becker, Student, Department of Mechanical and
Aerospace Engineering, University of Virginia.
David Sheffler, Professor, Department of Mechanical and
Aerospace Engineering, University of Virginia.