marks the beginning of a new era in flapping-wing flight. In
the past, we used a motor and gears to drive the flapping of
the wings. The wings had a set range of motion, and we could
only control the speed. Now, with a totally new paradigm, we
can exactly control the movement of each wing, using a powerful
actuator that is linked to an onboard computer. This system
mimics the muscles and brain of a real bird.
revolution had its earliest beginnings in 1985. Paul MacCready
was the legendary co-creator of the Gossamer Condor airplane,
along with Peter Lissaman. It was the first human-powered
airplane to fly a figure-eight course totaling one mile. In
1985, Paul MacCready’s company, Aerovironment, was funded
by the Smithsonian Institution to build a half-sized replica
of the giant pterosaur, Quetzalcoatlus northropi. It was featured
in an IMAX film called On the Wing. Even at half scale, the
robotic pterosaur had an 18 foot (5.5 meter) wingspan. It was the first ornithopter
to use an onboard computer to provide active stabilization
of an otherwise unstable aircraft. The wings were driven by
a powerful, custom-built servo, which provided variable flapping
amplitude and computerized control of the wing movements.
Although the QN replica was a milestone, the technology developed
in the $4 million project did not find its way into common
use in the ornithopter field.
the years, some people did begin to experiment with servo-powered
ornithopters. Initially, we used the same servos that are
used to move the control surfaces in radio-controlled airplanes.
This new task of flapping the wings would require far more
power. But we gave it a try anyway, because it seemed like
a much easier alternative than trying to build our own ornithopter
gearboxes completely from scratch.
At the time, we didn't
fully appreciate the other benefits that would follow.
servo-powered ornithopter attempts.
Bob Hoey's bird glider
Hoey was among the first to try servo-driven wing flapping.
He built a series of realistic bird-shaped gliders. The birds
were carried aloft by an RC plane, and then they glided down
from a height. Some of the models could
move their wings up and down. Hoey knew the servos he had
were not powerful enough for full flapping-wing flight, but
it gave the appearance.
Chaulet, around 1997, made an effort at powered flight using
hobby servos. He connected the servos to the wings using a
linkage, just as we had been doing with our gearboxes. He
built one RC bird with wings hinged at the body, and another
with hinges part way out on the wing. In this second model,
a single servo was used to flap both wings, and an additional,
smaller servo was provided for steering.
own first attempt was a simplified, lightweight design, with
one servo for each wing. To eliminate excess weight and friction,
the wing spars were attached directly to the servo arms. The
wing spars were secured with rubber o-rings, allowing their
easy removal. The servos were controlled directly by a small
RC receiver. I launched the model many times from a hill,
and this allowed me to gain some appreciation for the difficulty
of controlling wing movements in this manner. Despite considerable
effort, my first servo-powered ornithopter could not maintain
its altitude, and the servos would get quite hot in the attempt.
Dodd also made an early contribution, with the servo-powered
blimps that he began building in 1994. Freed of the restrictions
of gravity, Dodd (aided by John Piri of West Coast Blimps)
was able to fully develop a servo-powered drive system with
electronic control. That is, a programmable microchip onboard
the model does all the work of moving the servos back and
forth. It even incorporates steering movements from the RC
receiver. This avoids the need for repetitively moving the
RC control stick back and forth to flap the wings. It provides
much more accurate wing movements, for a more stable flight.
While he was working on these models, we still didn't have
any servos powerful enough to fly without helium.
Nathan Chronister's 2005 "Raven" ornithopter
with powerful, custom-built servos.
2005, I built a new type of ornithopter using a hybrid approach.
Called the “Raven” due to its onboard intelligence, it had
separate motors driving the left and right wings. Each motor
and gearbox was constructed in the conventional way, but I
incorporated a potentiometer into each wing hinge, to sense
the position of the wing. This information went to a microcontroller,
which could vary the amount of power to the two motors. The
positional feedback system amounted to a custom-built servo,
and it offered some benefits never before imagined in an ornithopter:
The two wings could be controlled independently, for a wide
range of different flight maneuvers. The wings provided all
flight control functions, and the tail was fixed. The wings
could be stopped at a favorable glide position. Unlike the
hobby servos available at the time, the dual motor drive system
provided easily enough power for flapping-wing flight. Unlike
hobby servos, the motors ran always in the same direction.
This was thought to provide better efficiency and power output.
However, this system does not allow varying the amplitude
of the wing flapping. It was also very elaborate to build.
after that, I learned that Hitec had introduced a whole new
type of servo. Their digital robot servo, HSR-5990TG, had
some advanced features that I thought might make it suitable
for driving the wings of a robotic bird. First, the new servo
used a better type of transistor, called MOSFETs. They switch
the motor cleanly off and on, with very little energy wasted
as heat. This would allow more efficient operation and a higher
power output. The new servos were also designed to operate
at a higher voltage, and they could produce far more power,
for their weight, than any previous hobby servo. It remained
to be seen if this would be enough power for flight. I
had the idea of launching a contest, to see who could come
up with the first ornithopter powered by these new servos.
I thought it would be a fascinating challenge, which could
bring new enthusiasts into the hobby. I thought it would stretch
people to come up with more efficient ornithopter designs,
to utilize what I still saw as a marginal power source. In
2011, I proposed the idea to Debra Cleghorn, at Model Airplane
News, and also to Hitec as a potential sponsor. Although there
was great interest in the concept, Cleghorn suggested that
I ought to build a test model first, to validate the concept.
I felt that it defeated the purpose of holding a contest to
see who could do it first. However, Hitec was kind enough
to donate a pair of the digital robot servos, to facilitate
the development of a prototype.
first ornithopter with the digital robot servos was not quite
successful. I tried to use a biplane configuration, based
on the classic Luna design. Many other researchers (and a
few toy companies) had adopted its scissor-wing design. The
balsa wing spars kept breaking, and even intact, the ornithopter
didn’t really want to fly. Biplanes were thought to require
about half as much power to fly, compared with the monoplane
type, but even so I decided that I would try again, with a
monoplane design. I had one ace in the hole, which was a new
wing design that could dramatically improve the performance
over conventional membrane wings, reduce the weight, and allow
it to fly with less power.
The S-1 Robotic Bird in flight.
S-1 Robotic Bird began its first test flight with a very simple
program and no radio control. Not sure what to expect, I set
it to flap the wings only five times and then go into a glide.
I waited for a perfectly calm morning, drove to the flying
field at the nearby elementary school, and stood there with
the model in my hand, trying to mentally prepare myself to
deliver a suitable hand launch without any way of knowing
how the model was going to behave. I connected the power and
positioned the giant bird for launch, while the programmed
time delay slowly elapsed. When the sleek robotic bird finally
took off, it darted forward from my hand and accelerated across
the field with startling vitality. After the planned five
wing beats, it went into a graceful glide and landed gently
in the grass.
found out the model would fly even better with some of Hitec’s
newer servos. The digital robot servo that I started with
is no longer made, and Hitec now sells a faster titanium-gear,
digital brushless servo that brings performance up to the
level of conventional RC ornithopters. Hitec has even incorporated
a regenerative braking system, which recaptures momentum at
the end of each wingstroke and makes the flight more efficient.
flights with radio control allowed me to vary the flapping
amplitude and experiment with some different steering methods.
Servo-powered ornithopters allow multiple steering methods,
without any additional servos, without any mechanical linkage,
and without any separate control surfaces. With only the two
servos that flap the wings, you can have four or more independent
control functions. You may also enjoy the challenge of experimenting
with various wing movements to see what kind of aerobatic
maneuvers you can perform. I am excited to get
this new platform into the hands of some other hobbyists,
and see what ideas they come up with. I created a new online
forum, accessible from BirdKit.com,
where people can share programming ideas and code for robotic
birds. Other than the programming, I didn’t have to make any
major changes between the original prototype and the production
model kit. This shows the versatility of these new servo-powered
ornithopters. Keep an eye out for updates, as significant
new developments are likely to unfold at a rapid pace.