High Speed Photography and Computational Fluid Dynamics may allow Engineers to Emulate Nature in Flying Micro-Robots.
Flying Micro-Robots offer multiple advantages to the military and law enforcement. Low observability in the visual, aural and radar domains, autonomous operation and portability are all highly desirable attributes of bio-mechanically inspired micro-robotic flight.Now, a bit over a century into the realm of powered flight, the door to natural flight and micro-robot duplication of bio-mechanical processes is opening a bit thanks to fascinating research into the flight dynamics of Dragonflies, Bumblebees and the Desert Locust recently published by researchers in the US, UK and Australia.
Using high speed cameras, a wind tunnel and advanced Computational Fluid Dynamics (CFD) modeling, Zoologists and Fluid Dynamics experts have made some fundamental discoveries. And as it turns out, there’s no best way to achieve flapping flight because the bio-mechanics are governed by why the insect is flying in the first place.
Consider the Dragonfly.
It’s fast, can change direction seemingly without effort, and can hover in a more stable manner than a Hummingbird. In fact, the beastie can even fly backwards! How can it do these things, and just as important from a research standpoint, why does it do these things?
The why is to help it capture prey and avoid becoming prey. Less capable flyers, like mosquitoes and other small insects, are no match for its superior flying skills. And few birds, the Dragonfly’s principle enemy, can compete when up against an alert member of the order Odonata, suborder Epiprocta. When the Dragonfly sees doom approaching, it simply darts off to one side at the last moment, leaving the bird nothing but empty air where it thought its next meal would be located.
But it’s the how of the aerobatic performance that takes a little getting used to.
Z. Jane Wang is a professor of theoretical and applied mechanics at Cornell University who has done landmark research into the flight dynamics of the Dragonfly. Wang believes that in some cases the apparently wasted up and down wing motion used by insects may be superior to the linear methods used in conventional airfoils.
In an interview posted on Cornell’s Chronicle website, Wang had this to say about the core of her research: “Dragonflies have a very odd stroke. It’s an up-and-down stroke instead of a back-and-forth stroke,” she said. “Dragonflies are one of the most maneuverable insects, so if they’re doing that, they’re probably doing it for a reason. But what’s strange about this is the fact that they’re actually pushing down first in the lift.
Figure 1: Note the positions of the Dragonfly’s four wings. The two front wings are in a stable position in this still frame and appear to be generating equal lift. The two back wings however do not match the position of the front wings nor do they match positions relative to one another. It appears that this Dragonfly is in the process of executing a lateral maneuver from a hover by inducing drag with the rear wings in unequal amounts. The resulting complex vortices can only be analyzed using Computational Fluid Dynamics.
“An airfoil uses aerodynamic lift to carry its weight. But the dragonfly uses a lot of aerodynamic drag to carry its weight. That is weird, because with airplanes you always think about minimizing drag. You never think about using drag.”
When the Dragonfly is moving forward, the two sets of wings are flapped in phase, generating lift and directional momentum.
When hovering however, the rear wings are intentionally flapped out of phase, with both the angle of attack and relative position of the front wings, which generates drag preventing motion relative to the earth but still generating lift. By decoupling one rear wing’s movement from the other, the insect can also move side to side.
All of these maneuvers require understanding of the vortices involved, because the vortices cancel or enhance lift as they interact, so a flexible CFD toolset is required.
But if these motions can be duplicated mechanically, they could give a micro-robot the ability to move rapidly from one location to another and then hover for purposes of surveillance. All the while the micro-robotic mini-spy would be generating an extremely small visual and Radar footprint plus an equally small noise signature.
Capabilities like these are of great interest to the military and law enforcement.
Forget what you’ve heard: “Bumblebees Can To Fly.”
Almost everyone even remotely interested in flight has heard the statement “Bumblebees shouldn’t be able to fly”. The belief is based on calculations using aerodynamic theory as it existed in 1918. These theories suggested that bumblebee wings were too small to create sufficient lift but CFD, and direct observation using stop motion photography, has shown the theories to be incorrect.
Richard Bomphrey and Adrian Thomas, Zoology professors at Oxford University in England, have learned that brute force is enough to overcome all sorts of aerodynamic taboos associated with levitating a large object using small lifting surfaces.
Bomphrey and Thomas determined that the Bumblebee’s huge thorax, with its intrinsic muscle mass, plus energy-rich nectar used for fuel, make up for an inefficient flying style.
Thomas says: “A bumblebee is a tanker-truck. Its job is to transport nectar and pollen back to the hive. Efficiency is unlikely to be important for that way of life.”
Figure 2: As this Bumblebee approaches a flower, the left wing has completed its downward motion and is rebounding upward. At the same time, the right wing is still moving downward. The airflow generated by the insect’s wings never meets as it does at the rear of an aeronautically efficient wing. This shortcoming makes the insect’s style of flight even less efficient. The lesson for engineers is to disregard the Bumblebee’s flight style unless energy is available in very large quantities.
The Oxford researchers found that it’s as if the insect is split in half. Not only do its left and right wings flap independently but the airflow around them never joins up to help it slip through the air more easily. Bomphrey says that an extreme aerodynamic separation between left and right like this is very unusual and adds to the insect’s inefficiency. But the Bumblebee’s inefficiencies are not without precedent.
Every time NASA launches the Space Shuttle, enormous quantities of energy are expended for a relatively small payload. Like the Bumblebee, brute force overcomes a lack of elegance but accomplishes a purpose. The Bumblebee’s bio-mechanics may eventually turn out to be useful in the robotic world but for those trying to create an efficiency leap by use of bio-mechanical flight methods, better just give the Bumblebee an award for unusual design approach and move on.
The Year of the Locust.
Figure 3. The Much Feared and Super Efficient Desert Locust.
Ever since Biblical times, hoards of Desert Locusts have flown intercontinental distances non-stop for one purpose: to strip the vegetation bare, convert it to energy, and move on to the next victim. Engineers working on micro-robots intended to stay aloft for long periods should be interested in just published research into the Locust’s bio-mechanics.
John Young, a scientist with the School of Engineering and Information Technology, University of New South Wales, Australia, teamed up with Professors Bomphrey and Thomas from Oxford to try and unlock the Locust’s secrets and they feel they’ve succeeded.
In their research paper, “Details of Insect Wing Design and Deformation Enhance Aerodynamic Function and Flight Efficiency”, the men report how they first mapped the wing surface geometry of the Locust using Gambit 2.4.6, a commercial meshing tool. 6636 cells were mapped on the forewing, 11740 cells were mapped on the hindwing and the body was divided into 5878 individual points. Kinematics were obtained using 4 high speed digital cameras recording the deforming surface topography of the wings while the Locust was tethered in a wind tunnel at its optimum long distance flight angle.
The unsteady, incompressible, Navier Stokes equations were solved using another commercial software package, Fluent 6.2.16. A dynamic mesh feature was used to create a moving boundary layer mesh so that wing movement, microscopic wing surface deformations and body interaction could be tracked throughout the wing beat cycles. When the simulations were run, they closely matched results obtained by injecting smoke into the wind tunnel’s air stream as the Locust flapped its wings.
Figure 4. A tethered Desert Locust is subjected to smoke derived from heated Johnson’s Baby Oil. The smoke was injected into a 3.3 ms-1 wind velocity generated by the wind tunnel. The body angle of the insect was set to 9°, the optimum angle for long distance travel. These data, obtained with a high speed Photron APX camera, validated the numerical data developed using the CFD tool.
The researcher’s have achieved full fidelity modeling of the complex structure and movement of the Locust’s wings for the first time. Using Gambit and Fluent CFD tools, Young, Bomphrey and Thomas proved that the Locust has mastered the art of minimizing airflow separation on the wing’s leading edge and controlling vortices on the trailing edge - twin efforts necessary for long distance, low energy flight. The Locust’s dynamic changing of the angle of attack during the wing beats seems to be another key to success.
But Young, Bomphrey and Thomas have found something else: the multitude of veins and individual cells that make up the wing structure are being deformed during the flapping of the wings. The types and magnitude of deformation depends on the task being executed.
The resulting changes in the shape and smoothness of the lifting surface have a direct effect on the efficiency of the airfoil and represent one of the most difficult challenges for engineers trying to duplicate the Locust’s amazing efficiency.
In the research paper, Young offers this advice to engineers. “Simple heaving or flapping of flat plates can generate high lift and stable vortices, but designing robust, lightweight wings that can also support efficient attached flow aerodynamics is likely to be much more difficult. Our results show that the secret in doing so lies in building a wing that undergoes appropriate aeroelastic deformation through the course of each wing beat.
“Our CFD simulations demonstrate that time-varying wing twist and camber are essential to maintenance of attached flow. Implementing such tailored deformations in an engineered system is a difficult problem and may demand an evolutionarily optimized solution in order even to approach the elegance of an insect.”
The Road Less Traveled.
There’s a good reason why powered human flight has evolved the way it has: It’s extremely tough to emulate natural flight. In addition to Professor Young’s observations, there are several other challenges involved.
First, the stresses involved in accelerating, stopping and reversing the motion of a mechanical wing will cause a great deal of wear on the components. In a bio-mechanical system, the body simply replaces the damaged cells. Also, mechanical actuators eventually wear out but a damaged muscle will simply heal.
Another difficult to duplicate attribute is adaptability. An animal or insect can learn how to fly using different flight dynamics in the event of injury provided the injury isn’t extensive. None of these self repair methods or adaptations are presently available for micro-robots.
There’s yet another aspect to consider which seems paradoxical at first glance: The lifetime for many insects is measured in weeks which is one reason that evolution has selected the bio-mechanical methodologies it has. The paradox is the MTBF of an insect’s component parts is excellent because it’s limited by the finite lifetime of the creature itself. What will be considered an acceptable lifetime for a mechanical insect? Will it also be measured in weeks?
Then there’s the control side of the equation: Because there are twisting motions, linear motions and deforming motions that vary with wing position and task, some sort of feedback network would have to be integrated into the flight surfaces and these data would then have to be assimilated by computers programmed to provide three dimensional wing geometry control as part of a complex feedback loop.
All of this increases weight, and since the first use for this technology would appear to be flying micro-robots used for surveillance, operational weight adds significant complications because of the need for a payload that would support a camera and a signal transmission system.
At the end of the day though, federal and local governments, as well as private industry, are very interested in the potential of bio-mechanics as a way to implement flying Micro-Robots. As a result, large amounts of money will likely be made available to solve these problems.
John Loughmiller is an EE, Commercial Pilot, Flight Instructor and a Lead Safety Team Representative for the FAA.
He can be reached at firstname.lastname@example.org.