Untethered Hovering Flapping - Am MIT spezialisierte KI-Forschung

Untethered Hovering Flapping
Flight of a 3D-Printed
Mechanical Insect

Charles Richter**
Cornell University

Hod Lipson*,**
Cornell University

Schlüsselwörter
3D printing, ornithopter, micro air vehicle,
untethered hovering, insect flight

A version of this paper with color figures is
available online at http://dx.doi.org/10.1162/
artl_a_00020. Subscription required.

Abstract This project focuses on developing a flapping-wing
hovering insect using 3D-printed wings and mechanical parts. Der
use of 3D printing technology has greatly expanded the possibilities
for wing design, allowing wing shapes to replicate those of real insects
or virtually any other shape. It has also reduced the time of a wing
design cycle to a matter of minutes. An ornithopter with a mass of
3.89 g has been constructed using the 3D printing technique and
has demonstrated an 85-s passively stable untethered hovering flight.
This flight exhibits the functional utility of printed materials for
flapping-wing experimentation and ornithopter construction and
for understanding the mechanical principles underlying insect flight
and control.

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

:
/
/

D
ich
R
e
C
T
.

ich
T
.

e
D
u
A
R
T
l
/

A
R
T
ich
C
e
–
P
D

F
/

1
7
2
7
3
1
6
6
2
8
2
3
A
R
T
l

_
A
_
0
0
0
2
0
P
D

B
j
G
u
e
S
T

Ö
N
0
7
S
e
P
e
M
B
e
R
2
0
2
3

1 Einführung

Hovering flapping flight of insects and birds has long fascinated scientists and engineers, but only in
the last decade has it been successfully demonstrated by man-made flying machines. Unlike forward
flight, hovering flapping flight poses several special challenges. Erste, there has yet to emerge an
established body of theoretical and experimental work on the unsteady aerodynamics of flapping
wing flight for the purposes of wing design. Zweite, hovering flapping flight of insects and birds
is generally unstable and requires a sophisticated solution to maintain an upright flying position [15,
16]. Dritte, the energy density of batteries was insufficient for the power demands of hovering flight
until small lithium-based batteries became widely available. Jedoch, with the improvement of elec-
trical power solutions, a number of successful hovering ornithopters have been developed with a
variety of wing designs. This project utilizes existing solutions to the power and stability problems
and uses 3D printing as a novel approach to designing and manufacturing the key aerodynamic
component: the wings.

Thus far, producing effective flapping wings for research and ornithopter construction has been a
time-consuming and delicate process taking days or longer to complete. The 3D printing technique
allows wings to be produced in a matter of minutes, dramatically reducing the time of each design
Zyklus. Overcoming this barrier to experimentation will allow a comprehensive study of lift produc-
tion for a wide variety of wing shapes, including those replicating real insect wings.

* Contact author.
** Computational Synthesis Laboratory, 239 Upson Hall, Sibley School of Mechanical & Aerospace Engineering, Cornell University, Ithaca,
New York 14853. Email: car45@cornell.edu (C.R.); hod.lipson@cornell.edu (H.L.)

Artificial Life 17: 73–86 (2011)

C. Richter and H. Lipson

Untethered Hovering Flapping Flight of a 3D-Printed Mechanical Insect

A comprehensive understanding of flapping-wing aerodynamics and hovering flight will become
increasingly important as ornithopters shrink to the scale of real insects, where some advantages of
flapping-wing flight are realized [7]. These advantages include efficiency and maneuverability im-
provements over fixed and rotary wing aircraft at low Reynolds numbers as well as the suitability
of microscale actuators to producing vibrating motion for flapping rather than rotary motion for
traditional propellers [12, 20]. Maneuverable, low-power micro air vehicles have a wide range of
applications, including mapping, Überwachung, and search-and-rescue operations, where these proper-
ties of small size and ability to maneuver in tight spaces are vital, and those in thin extraterrestrial
atmospheres, where low Reynolds numbers occur [11]. Micro air vehicles also present a challenging
synthesis of many areas of engineering, including materials, actuators, electronics, Kontrolle, vision,
and guidance [8, 9]. This project has demonstrated the viability of 3D-printed aerodynamic compo-
nents for experimentation and for use in a real ornithopter on the size scale of the smallest current
designs (Figur 1).

1.1 Review of Existing Work
The existing work that has influenced this project includes a variety of successful ornithopter designs
and some research on the dynamics and control of insect flight (Tisch 1). This project is a continua-
tion of an earlier ornithopter design project by Floris van Breugel of the Cornell Computational
Synthesis Laboratory. Van Breugelʼs design used four motors to drive eight wings and featured pas-
sively stable flight dynamics using a set of damping sails above and below the body of the aircraft.
This model had a mass of 24 g and demonstrated stable hovering flight of over 30 s in 2007. Broad
goals for the current project were to achieve a comparable flight time using this system of passive
stability in a vehicle under 10 G.

Several other successful designs currently exist, including the series of DelFly ornithopters, welche
are radio controlled using tail configurations resembling fixed-wing aircraft, and the AeroVironment
Nano Air Vehicle, which achieves control using active wing control. The Harvard Microrobotics Lab-
oratory has also produced ornithopters weighing 60 mg using piezoelectric actuators and insectlike
passive wing pitching, but requiring a tether for power and stability.

There have also been recent developments in the understanding of insect flight [2, 6, 13, 18].
These studies have explored one mechanism of passive wing deflection in insect flight that is essen-
tial to the simplicity of some ornithopter designs. They have shown that some insect wings deflect
to an angle of incidence of 45°, which is thought to be optimal for lift production of a flat-plate
wing. These findings have also given rise to hypotheses explaining forward thrust, flight maneuvers,

Figur 1. 3D-printed elements of flapping-hovering insect.

Artificial Life Volume 17, Nummer 2

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

:
/
/

D
ich
R
e
C
T
.

ich
T
.

e
D
u
A
R
T
l
/

A
R
T
ich
C
e
–
P
D

F
/

1
7
2
7
3
1
6
6
2
8
2
3
A
R
T
l

_
A
_
0
0
0
2
0
P
D

B
j
G
u
e
S
T

Ö
N
0
7
S
e
P
e
M
B
e
R
2
0
2
3

C. Richter and H. Lipson

Untethered Hovering Flapping Flight of a 3D-Printed Mechanical Insect

Tisch 1. Characteristics of existing ornithopter designs.

Design

Year

Mass (G)

Span (cm)

Wings

Hover time (S)

Features

Mentor [21]

2002

580

DelFly II [5]

2006

16.07

van Breugel [17]

2007

24.2

Chronister [3]

2007

3.3

Holz [19]

2007

0.060

DelFly Micro [5]

2008

3.07

NAV [1]

2009

10 (est.)

7.5 (est.)

Richter (this work)

2010

3.89

14.3

>60

480

Nitromethane fuel

Camera, R/C

Passively stable

Unknown

R/C

N/A

Piezoelectric power

Camera, R/C

Active wing pitching

3D printed parts, wings

and disturbance rejection, and experiments have been designed to examine these hypotheses using
the ornithopter as a test bed.

One primary goal of this project was to produce a hovering ornithopter with as many 3D-printed
components as possible. An Objet EDEN260V printer and the Objet FullCure 720 material were
used to produce all printed components. This material costs roughly 0.22 USD per gram, und das
EDEN 260V prints with a resolution of 42 Am on the x and y axes and 16 Am on the z axis. At first,
only the fuselage, hinges, and pushrods were printed; Jedoch, a method of printing entire one-piece
wings was soon developed.

First attempts at wing construction were aimed at recreating the wings of the van Breugel de-
sign, using a carbon fiber rod as the main strut, polyethylene terephthalate (HAUSTIER) stiffening ribs, Und
a Mylar film wing surface. Two examples of this early printed type can be seen in the upper left
corner of Figure 2. The carbon fiber rod was to extend out of a 3D-printed hinge, but after several
design iterations, the hinge, strut, and stiffening ribs were combined into a single printed piece.
When further experimentation revealed that a durable thin film could be printed using only two

Figur 2. A variety of wing shapes for experimentation.

Artificial Life Volume 17, Nummer 2

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

:
/
/

D
ich
R
e
C
T
.

ich
T
.

e
D
u
A
R
T
l
/

A
R
T
ich
C
e
–
P
D

F
/

1
7
2
7
3
1
6
6
2
8
2
3
A
R
T
l

_
A
_
0
0
0
2
0
P
D

B
j
G
u
e
S
T

Ö
N
0
7
S
e
P
e
M
B
e
R
2
0
2
3

C. Richter and H. Lipson

Untethered Hovering Flapping Flight of a 3D-Printed Mechanical Insect

layers of printed material, this film was used instead of Mylar as the wing surface, and the first one-
piece printed wings were made. Figur 2 shows many conventional and biologically inspired printed
wings.

1.2 Printed-Wing Construction
The printed wings of the ornithopter are composed of three functional elements: the central beam,
the surrounding frame, and the thin film wing surface. Figur 3 shows the parts of the dual wing
used in the full ornithopter design. The central beam is the most rigid portion of the wing and
contains the pivot point as well as the attachment holes for the connecting rods. Whereas some
designs require a bushing or dedicated hinge, 3D printing allows the hinge to be incorporated into
the main beam design. Außerdem, the FullCure 720 material features relatively low friction against
the stainless steel 0.5-mm piano wire hinge pins when lubricated with a drop of medium-viscosity
oil. The holes for the pivot points were designed with a 0.6-mm diameter to provide an adequate
gap for low-friction operation. This technique eliminates the need for a heavy bushing or complex
assembly.

The outer frames of the wings are attached to the ends of the beam. The outer frames determine
the flexibility of the wings and the deflection properties during flapping. The outer frames were
defined in the CAD model as lofted curves connecting circular cross sections. By varying the radius
of the circular cross sections at various points along the frame, the overall stiffness and flexibility
patterns of the wing could be tuned.

The thin wing surface is a flexible film that extends through the area inside the outer frame. Der
surface has a thickness of 40 Bin, which is achieved by depositing two layers of material. The ability
of the printer to print such a thin flexible film is the development that made a one-piece printed
wing possible. While it is possible to print a thinner film using a single layer, wings constructed with a
single-layer surface are extremely delicate and tend to tear upon vigorous flapping. Chamfers were
used to counter the tendency of the wing film to tear at points of discontinuous geometry, wie zum Beispiel
the edge where the film joins the frame.

One practical element of 3D printing technology is the use of a gelatinous material to support the
structure during printing. daher, removing the support material is an important step in the man-
ufacturing process, especially with delicate features such as the thin wingʼs surface. Common meth-
ods used to remove support material include dissolving it in sodium hydroxide and spraying it off
with pressurized water. Jedoch, both of these methods have limitations due to the delicacy of the
thin film. When a printed wing is soaked in liquid for any period of time, it tends to curl up or warp.
That can be partially corrected by pressing it flat and allowing it to dry. Jedoch, the moisture tends
to leave some permanent warping of the wing shape. The method of spraying pressurized water is
also difficult because extreme care must be taken to avoid tearing the wing film. Wieder, the moisture
tends to warp the wing shape. The best method thus far has been to place the wing on a clean
surface with some elasticity, such as a dense rubber mat, and scrape the support material away using

Figur 3. Parts of the one-piece printed wing.

Artificial Life Volume 17, Nummer 2

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

:
/
/

D
ich
R
e
C
T
.

ich
T
.

e
D
u
A
R
T
l
/

A
R
T
ich
C
e
–
P
D

F
/

1
7
2
7
3
1
6
6
2
8
2
3
A
R
T
l

_
A
_
0
0
0
2
0
P
D

B
j
G
u
e
S
T

Ö
N
0
7
S
e
P
e
M
B
e
R
2
0
2
3

C. Richter and H. Lipson

Untethered Hovering Flapping Flight of a 3D-Printed Mechanical Insect

Figur 4. Experimental test setup on the lab scale.

a dull blade. Any residual material can be removed by wiping with a cloth moistened with water or
rubbing alcohol. This is the fastest and most successful method for removing support material from
the thin wing film.

1.3 Wing Design
At the beginning of the project, the wing design process focused on narrowing the vast design space
to a size scale that was appropriate for the motors available and desired weight of the vehicle. Während
initial testing, key wing design features were identified that helped produce the ideal shapes and
deflections when flapping. Testing of a wide variety of wing shapes, sizes, and structures was carried
out by powering them with a small DC gear motor using a DC power source. The lift of each wing
was measured using a custom attachment for a digital lab scale, and flapping behavior was analyzed
using a high-speed camera capturing 1000 frames per second. Figuren 4 Und 5 show the experimental
apparatus.

The wing size partially determines several important variables, including the mass and surface
Bereich, which in turn determine how fast the wings can flap for a given power input. For the motor
chosen for this project (a GM15 gear motor available from Solarbotics.com with 25:1 gear reduc-
tion) and the power expected from a pair of lithium-polymer batteries (7.4 V, 200 mA), the best-
performing single wing of all wing designs tested had a length of 80 mm and a maximum chord of
30 mm. The overall weight of the wing was approximately 0.3 G, and the thickness of the wing film

Figur 5. Close-up of test setup mechanism.

Artificial Life Volume 17, Nummer 2

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

:
/
/

D
ich
R
e
C
T
.

ich
T
.

e
D
u
A
R
T
l
/

A
R
T
ich
C
e
–
P
D

F
/

1
7
2
7
3
1
6
6
2
8
2
3
A
R
T
l

_
A
_
0
0
0
2
0
P
D

B
j
G
u
e
S
T

Ö
N
0
7
S
e
P
e
M
B
e
R
2
0
2
3

C. Richter and H. Lipson

Untethered Hovering Flapping Flight of a 3D-Printed Mechanical Insect

Figur 6. The most successful wing design during testing.

War 40 Bin. This wing flapped at approximately 30 Hz through an angle of 110° and produced a
maximum lift of 2.92 G. This wing design is shown in Figure 6.

The wing structure is important to proper deflection and wing shape during flapping. For max-
imum lift, the wing should deflect to an angle of attack of roughly 45° at the middle of the stroke.
This angle of attack can be tuned by adjusting the flexibility of the main wing strut and the ribs that
stiffen the interior of the wing. Thus far, successful wing designs have been created with and without
wing ribs.

One major problem associated with simple deflecting wings is that they do not deflect as flat
plates. Stattdessen, the leading edge tends to remain vertical rather than flexing torsionally, während die
wing surface bends away underneath it. This behavior creates an inverted camber shape that is un-
desirable. Several methods were explored to overcome this problem. The most effective solution
was to extend the wing frame all the way around the tip of the wing. This design forced the leading
edge to twist when the wing deflected, thus maintaining a roughly continuous slope across the chord
of the wing near the tip. Mit anderen Worten, the tip of the wing behaved more like a flat plate, mit dem
entire wing deflecting to the proper angle rather than just the lower half.

Wing ribs have also been used to control the deflection patterns and add stiffness in certain di-
rections. Various rib designs were tested featuring rectilinear patterns as well as curved patterns
inspired by the wings of dragonflies and other insects. Jedoch, the current design does not feature
stiffening ribs. Figur 7 shows a top-down view of a wing deflecting during flapping tests on the
experimental setup. This general wing design, while not optimal, was deemed satisfactory for use in
the challenge of building a full ornithopter using 3D-printed wings. A new double-ended version of
this wing shape was produced for use in the full ornithopter.

1.4 Full Ornithopter Design
Once a satisfactory wing design was obtained, it was implemented in the four-wing vehicle. Der
wing chosen for this purpose was the ribless design that produced the greatest lift. A fuselage
was designed to hold the motor, crank, and wing hinge. Care was taken to place the motor as close
as possible to the wing pivot point to center the mass.

Figur 7. Flash photo showing deflection while flapping.

Artificial Life Volume 17, Nummer 2

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

:
/
/

D
ich
R
e
C
T
.

ich
T
.

e
D
u
A
R
T
l
/

A
R
T
ich
C
e
–
P
D

F
/

1
7
2
7
3
1
6
6
2
8
2
3
A
R
T
l

_
A
_
0
0
0
2
0
P
D

B
j
G
u
e
S
T

Ö
N
0
7
S
e
P
e
M
B
e
R
2
0
2
3

C. Richter and H. Lipson

Untethered Hovering Flapping Flight of a 3D-Printed Mechanical Insect

Figur 8. Top view of ornithopter mechanism showing offset-crank geometry.

The wings are driven by a crankshaft connected to the motorʼs gearbox. In order to drive the
wings in a roughly symmetrical motion, the crankshaft includes two attachment points for the
connecting rods powering the left and right wings. These two attachment points are roughly 30°
out of phase from each other to compensate for the asymmetry of the crank position at any given
point in the stroke. Figuren 8 Und 9 show this offset-crank mechanism, which is similar to the DelFly I
Design [4] and many toy ornithopters.

The ornithopter was tested first using a DC power source and a fishing line tether to verify proper
operation of the crank mechanism and proper flapping behavior of the wings. The crank is designed
to flap each of the four wings through roughly 80°, and when the flexibility of the wings is
enthalten, this angle is enough to allow the wings to clap and fling at the end of each stroke. Der

Figur 9. Large view of ornithopter mechanism.

Artificial Life Volume 17, Nummer 2

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

:
/
/

D
ich
R
e
C
T
.

ich
T
.

e
D
u
A
R
T
l
/

A
R
T
ich
C
e
–
P
D

F
/

1
7
2
7
3
1
6
6
2
8
2
3
A
R
T
l

_
A
_
0
0
0
2
0
P
D

B
j
G
u
e
S
T

Ö
N
0
7
S
e
P
e
M
B
e
R
2
0
2
3

C. Richter and H. Lipson

Untethered Hovering Flapping Flight of a 3D-Printed Mechanical Insect

clap-and-fling phenomenon may aid in lift production [10]. Figur 10 shows a photo of a tethered
flight test showing ideal wing deflection of roughly 45°. In this test configuration, the ornithopter
was able to lift up to 1.5 g of payload, which is roughly equivalent to the mass of batteries required
for flight.

Once the ornithopter was able to support a payload while flying on the tether, it was outfitted
with batteries, and untethered flight tests began. Two 10-mA h lithium-polymer batteries were used
to power the motor and were attached on the opposite side of the motor to balance the mass. Der
other feature required for untethered flight is a set of thin foam damping sails attached to a thin
carbon fiber rod above and below the fuselage to maintain an upright flying position. This method
of achieving passive stability was developed by van Breugel and is replicated here [17].

1.5 Passive Stability
The stabilizing sails shown in Figure 11 are designed to maintain stability and keep the ornithopter
in an upright orientation. Without sails, the ornithopter tends to tip over, causing a horizontal accel-
eration and loss of upward lift. Jedoch, when the sails are attached, the larger top sail provides a
righting force to counter the tendency to tip over, while the bottom sail dampens any pendulum-like
oscillations. If launched upside down, the ornithopter will right itself, demonstrating the robustness
of the design.

1.6 Lift and Power Characterization
The four-wing configuration was evaluated in a series of experiments to characterize lift and power
Leistung. A test apparatus was designed to replicate the exact geometry and kinematics of the
ornithopter fuselage; Jedoch, a more powerful motor was selected in order to flap the wings at a
steady frequency for extended periods without overheating. For this purpose, the Mabuchi FF-050
motor and BaneBots 11:1 gearbox were used. The four-wing test setup was mounted to a laboratory
scale to measure the average lift force and is shown in Figure 12.

Five identical sets of wings were 3D-printed, and each set was tested individually. During each
test, the wings were flapped at a range of speeds while voltage, current, frequency, and mean lift
force were recorded. Frequency was measured using a stroboscope, while voltage and current were
measured using multimeters. The power required to drive the motor and linkage mechanism alone
(without wings) was also recorded over a range of frequencies. At each frequency, this value was
subtracted from the total power supplied to the system to give a meaningful measurement of the
inertial and aerodynamic power demands of flapping lift production.

Figur 10. Wing deflection in a tethered flight test.

Artificial Life Volume 17, Nummer 2

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

:
/
/

D
ich
R
e
C
T
.

ich
T
.

e
D
u
A
R
T
l
/

A
R
T
ich
C
e
–
P
D

F
/

1
7
2
7
3
1
6
6
2
8
2
3
A
R
T
l

_
A
_
0
0
0
2
0
P
D

B
j
G
u
e
S
T

Ö
N
0
7
S
e
P
e
M
B
e
R
2
0
2
3

C. Richter and H. Lipson

Untethered Hovering Flapping Flight of a 3D-Printed Mechanical Insect

Figur 11. Final configuration of complete ornithopter.

The plots in Figures 13 Und 14 show the lift force as a function of frequency and flapping power.
The dashed line represents the mass of the complete ornithopter, broken down in Figure 15, welche
is the lift required for hovering. The standard deviations in these experiments, represented by the
error bars in Figures 13 Und 14, are very small. The mean standard deviation of lift force across all
frequencies of flapping in these experiments was just 1.53%, indicating excellent consistency among

Figur 12. Four-wing test stand precisely replicating the geometry and kinematics of the real ornithopter fuselage for the
purpose of characterizing lift and power performance of the four-wing configuration.

Artificial Life Volume 17, Nummer 2

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

:
/
/

D
ich
R
e
C
T
.

ich
T
.

e
D
u
A
R
T
l
/

A
R
T
ich
C
e
–
P
D

F
/

1
7
2
7
3
1
6
6
2
8
2
3
A
R
T
l

_
A
_
0
0
0
2
0
P
D

B
j
G
u
e
S
T

Ö
N
0
7
S
e
P
e
M
B
e
R
2
0
2
3

C. Richter and H. Lipson

Untethered Hovering Flapping Flight of a 3D-Printed Mechanical Insect

Figur 13. Average lift produced by the four-winged ornithopter configuration as a function of flapping frequency in five
independent trials with error bars showing standard deviation.

3D-printed copies of the same wing design. This consistency also reflects the precision available for
tuning the wing structure, which makes 3D printing a versatile and powerful technique for wing
Experimentieren.

2 Conclusions

This project has yielded several significant results thus far. Erste, wing tests and the hovering dem-
onstration have validated the concept of a printed ornithopter. This method of construction has greatly
accelerated the design cycle, since a set of wings can be printed in less than 30 min, and a complete set
of ornithopter parts can be printed in 60 min. Daher, several design iterations can be tested per day.

Figur 14. Average lift as a function of flapping power, defined as the total power applied to the motor minus the power
required to drive the motor and linkage mechanism alone at the same frequency, in five independent trials, with error
bars indicating standard deviation.

Artificial Life Volume 17, Nummer 2

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

:
/
/

D
ich
R
e
C
T
.

ich
T
.

e
D
u
A
R
T
l
/

A
R
T
ich
C
e
–
P
D

F
/

1
7
2
7
3
1
6
6
2
8
2
3
A
R
T
l

_
A
_
0
0
0
2
0
P
D

B
j
G
u
e
S
T

Ö
N
0
7
S
e
P
e
M
B
e
R
2
0
2
3

C. Richter and H. Lipson

Untethered Hovering Flapping Flight of a 3D-Printed Mechanical Insect

Figur 15. Breakdown of total mass (3.89 G).

The Objet FullCure 720 material has some limitations, particularly in its mechanical properties. Es
is not as light or as stiff as carbon fiber or balsa wood, which are the main alternative options for
wing struts. daher, printed wings do not store as much energy when they flex, and energy is lost
to friction during each wing stroke. Different strut cross sections will be tested to improve the stiff-
ness per unit volume of material.

Other limitations of the 3D printed material include a tendency of thin wings to curl up after a
period of days, rendering them useless. This problem can be corrected by storing wings between flat
plates or in the pages of a book, though that requires disassembly. Thin wings also tend to develop
small tears after minutes of vigorous flapping. Jedoch, this problem can be partially prevented with
chamfered edges along the wing frame to avoid discontinuous geometry.

Experimentation with wing designs has begun to uncover some of the features and parameters of
successful wings for this size and power scale. The GM15 motor seems to be well matched to wings
that are approximately 80–100 mm long from base to tip with a chord length of 30–40 mm when it
is running at a power of 1.5 W (typical power consumption during flight). If the wing strut is ex-
tended further, then the drag of the wing acts along a longer lever arm, slowing down the rate of
flapping and reducing lift.

One very successful design feature is the wing frame that extends around the wingtip. This fea-
ture helps maintain a continuous wing slope at the tip of the wing and helps approximate the flat-
plate airfoil cross section of many hovering insects. The continuous wingtip frame was a design
borrowed from the structure of dragonfly wings, which exhibit ideal shape and deflection at the
wingtips. Gesamt, the use of 3D printing to create flexible wings that are aerodynamically functional
is the main accomplishment of this project and will be one area for future improvement. The com-
plete 3D-printed ornithopter is shown in Figures 16 Und 17 and is shown hovering in Figure 18.

3 Future Work

A long-term project utilizing a hovering ornithopter will be to test hypotheses of insect propulsion
and control. This project will be carried out by building wings with a nominal bias of several degrees

Artificial Life Volume 17, Nummer 2

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

:
/
/

D
ich
R
e
C
T
.

ich
T
.

e
D
u
A
R
T
l
/

A
R
T
ich
C
e
–
P
D

F
/

1
7
2
7
3
1
6
6
2
8
2
3
A
R
T
l

_
A
_
0
0
0
2
0
P
D

B
j
G
u
e
S
T

Ö
N
0
7
S
e
P
e
M
B
e
R
2
0
2
3

C. Richter and H. Lipson

Untethered Hovering Flapping Flight of a 3D-Printed Mechanical Insect

Figur 16. Final design with sails for passive stability.

built into the angle of incidence to produce forward thrust or turning maneuvers. If successful, diese
principles could form the basis of hovering ornithopter control.

Another project planned for the future is to perform a detailed study using 3D-printed wings
to develop analytical models predicting wing performance. The lift of many different wing de-
signs will be measured to identify relationships between the major variables involved in lift produc-
tion, such as wing length, chord, surface area, flapping frequency, and parameterized shape. Das
data will then be mined for analytical relationships using the Eureqa software [14]. These laws will
then be compared with current designs to evaluate the model and ultimately produce the best pos-
sible wings.

Endlich, another ornithopter will be designed using 3D-printed wings and other parts that is still

smaller and lighter and is composed of an even greater proportion of printed components.

Figur 17. Ornithopter mechanism close-up.

Artificial Life Volume 17, Nummer 2

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

:
/
/

D
ich
R
e
C
T
.

ich
T
.

e
D
u
A
R
T
l
/

A
R
T
ich
C
e
–
P
D

F
/

1
7
2
7
3
1
6
6
2
8
2
3
A
R
T
l

_
A
_
0
0
0
2
0
P
D

B
j
G
u
e
S
T

Ö
N
0
7
S
e
P
e
M
B
e
R
2
0
2
3

C. Richter and H. Lipson

Untethered Hovering Flapping Flight of a 3D-Printed Mechanical Insect

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

:
/
/

D
ich
R
e
C
T
.

ich
T
.

e
D
u
A
R
T
l
/

A
R
T
ich
C
e
–
P
D

F
/

1
7
2
7
3
1
6
6
2
8
2
3
A
R
T
l

_
A
_
0
0
0
2
0
P
D

B
j
G
u
e
S
T

Ö
N
0
7
S
e
P
e
M
B
e
R
2
0
2
3

Figur 18. Ornithopter taking flight and hovering in time-series images (A), (B), Und (C).

Artificial Life Volume 17, Nummer 2

C. Richter and H. Lipson

Untethered Hovering Flapping Flight of a 3D-Printed Mechanical Insect

Danksagungen
This work was supported in part by the U.S. Nationale Wissenschaftsstiftung (NSF) Grant ECCS
0941561 on Cyber-enabled Discovery and Innovation (CDI). We thank Leif Ristroph, Itai Cohen,
and Jane Wang for useful discussions.

Verweise
1. AeroVironment (2009). DARPA awards AeroVironment Phase II contract extension for nano air vehicle

development program. Available at http://www.avinc.com/downloads/NAVPRLongDARPAV4.doc.pdf
(accessed April 2010).

2. Bergou, A. J., Xu, S., & Wang, Z. J. (2007). Passive wing pitch reversal in insect flight. Zeitschrift für

Fluid Mechanics, 591, 321–337.

3. Chronister, N. (2010). Micro air vehicle ornithopters. Available at http://www.ornithopter.org/history.

mav.shtml (accessed April 2010).

4. de Croon, G. C. H. E., de Clerg, K. M. E., Ruijsink, R., Remes, B., & de Wagter, C. (2009). Design,
aerodynamics, and vision-based control of the DelFly. International Journal of Micro Air Vehicles, 1(2),
71–97.

5. DelFly. (2010). Available at http://www.delfly.nl/ (accessed April 2010).

6. Dickinson, M. H., Lehmann, F., & Sanjay, P. (1999). Wing rotation and the aerodynamic basis of insect

flight. Wissenschaft, 284, 1954–1960.

7. Ellington, C. P. (1999). The novel aerodynamics of insect flight: Applications to micro-air vehicles.

Journal of Experimental Biology, 202(23), 3439–3448.

8. Floreano, D., Zufferey, J. C., Srinivasan, M., & Ellington, C. (Hrsg.). (2010). Flying insects and robots.

Berlin: Springer.

9. Karpelson, M., Wei, G. Y., & Holz, R. J. (2008). A review of actuation and power electronics options for

flapping-wing robotic insects. In IEEE International Conference on Robotics and Automation.

10. Lehmann, F., Sane, S. P., & Dickinson, M. (2005). The aerodynamic effects of wing-wing interaction in

flapping insect wings. Journal of Experimental Biology, 208, 3075–3092.

11. Michelson, R., & Naqvi, M. (2003). Extraterrestrial flight (entomopter-based Mars surveyor ). In

von Karman Institute for Fluid Dynamics RTO/AVT lecture series on low Reynolds number aerodynamics on aircraft
including applications in emerging UAV technology. Brussels, Belgien.

12. Pesavento, U., & Wang, Z. J. (2009). Flapping wing flight can save aerodynamic power compared to

steady flight. Physical Review Letters, 103(11), 118102.

13. Ristoph, L., Berman, G. J., Bergou, A. J., Wang, Z. J., & Cohen, ICH. (2009). Automated hull reconstruction
motion tracking (HRMT) applied to sideways maneuvers of free-flying insects. Journal of Experimental
Biology, 212, 1324–1335.

14. Schmidt, M., & Lipson, H. (2009). Distilling free-form natural laws from experimental data. Wissenschaft,

324(5923), 81–85.

15. Sun, M., & Xiong, Y. (2005). Dynamic flight stability of a hovering bumblebee. Journal of Experimental

Biology, 208, 447–459.

16. Taylor, G. K., & Thomas, A. L. R. (2003). Dynamic flight stability in the desert locust Schistocerca gregaria.

Journal of Experimental Biology, 206, 2803–2829.

17. van Breugel, F., Regan, W., & Lipson, H. (2008). From insects to machines: A passively stable,

untethered flapping-hovering micro air vehicle. IEEE Robotics and Automation Magazine, 15, 68–74.

18. Wang, Z. J. (2005). Dissecting insect flight. Annual Review of Fluid Mechanics, 37, 183–210.

19. Holz, R. J. (2008). The first takeoff of a biologically-inspired at-scale robotic insect. IEEE Transactions

on Robotics, 24, 341–347.

20. Wald, M. ICH., Henderson, J. F., & Lock, G. D. (2001). Energy requirements for the flight of micro

air vehicles. Aeronautical Journal, 105(1045), 135–149.

21. Zdunich, P., Bilyk, D., Mac Master, M., & Loewen, D. (2007). Development and testing of the

Mentor flapping-wing micro air vehicle. Journal of Aircraft, 44, 1701–1711.

Artificial Life Volume 17, Nummer 2

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

:
/
/

D
ich
R
e
C
T
.

ich
T
.

e
D
u
A
R
T
l
/

A
R
T
ich
C
e
–
P
D

F
/

1
7
2
7
3
1
6
6
2
8
2
3
A
R
T
l

_
A
_
0
0
0
2
0
P
D

B
j
G
u
e
S
T

Ö
N
0
7
S
e
P
e
M
B
e
R
2
0
2
3 Untethered Hovering Flapping image

PDF Herunterladen