Quintessential Robotic Research 
A FLAPPING WING, HOVERING AND FLYING ORNITHOPTER BAT ROBOT

"ornithopter" from "orni"=bird and "thopter"= machine. 
A machine that is held aloft and propelled by wing movements. 
In search of the Holy Grail to build a machine that can fly and hover like a bird or bat. 
Here are some links:- 

http://www.popsci.com/popsci/bown/article/0,16106,388615,00.html

http://www.ucmp.berkeley.edu/vertebrates/flight/bats.html 

http://www.ornithopter.org/bb.html

http://www.ornithopter.org/toporov/toporov.html

http://www.nurseminerva.co.uk/adapt/flapping1.htm

http://www.intercept-technologies.com 

http://home.t-online.de/home/LilienthalMuseum/e213.htm#1

http://www.space.com/scienceastronomy/solarsystem/mars_flapper_011205-1.html

http://www.eng.fsu.edu/~dommelen/courses/eml5935/00/topics/11_28/sld001.htm

http://www.aerovironment.com/area-aircraft/unmanned.html

http://www.chez.com/truefly/Images/ctrflya2.htm

Hey do you know what is crazy interesting about the control of a bat robot? It's just like walking around on an asteroid that has only ten times your own mass!! To explain what I mean we need to remember how we can walk around on the earth. We do it by pushing the earth backwards. Overall momentum of the walker and the earth is conserved but we lose out a little by increasing the angular kinetic energy of the earth by a tiny little bit which means that there is an inefficiency in the operation. You can show using mathematics that we impart negligibly small angular kinetic energy into the earth and almost all the energy we use in moving around goes into our own body mass. However if we walk around an asteroid which is only ten times our mass then we can't increase our kinetic or potential energy without a significant amount going into the asteroid. You can think of the bat's body as the asteroid and the wings as us. When the bat flaps its wings it can only accelerate them by obtaining a reaction from its body. The only way the body can give a reaction is for itself to be accelerated in opposition. So it appears that designing a bat robot is like space engineering. The aerodynamic forces are another issue. 

At present we have built underwater robots and ground based robots so the next hurdle is flying robots. We are trying to think of a different approach to flying robots. The robot needs to hover and manoeuvre on the spot but not a helicopter and not an airship. It also needs to fly like a conventional aeroplane. Maybe the tilting wing Osprey system or the vectored thrust Harrier? But what's the point in following everybody else, why not try and do something different for a change and see what comes out. 

The bird appears to have three types of flight. Hovering, transitional and fully developed. This is not in the books as far as I can see. The books seem only to go half way in explaining how birds fly. You see diagrams of a wing on an upstroke or downstroke with a vector representing the apparent wind but you don't get a complete picture. To be fair I don't think you can get a true idea of flapping wing flight until you make a machine that does the same thing. Only then do you fully appreciate the problems, then solve them, then understand them. Does the flapping wing in hovering flight (not humming bird hovering flight which does not use retraction or protraction of the wing) arrest the wing periphery vortex from the previous downstroke and so give additional lift? Is the lift from a flapping wing in hovering flight proportional to the flapping frequency squared? How can you flap at high speed? What is the purpose of the wing hingeing at the wrist? Why does the bird make it seem so effortless to fly? What a beautiful well-oiled machine. Academics go on as usual simulating and writing papers and filling the pages full of seemingly clever equations but I still don't get it so it's back to paper and pencil and making real machines and figure it out for yourself. (What is required is a new approach to Product Design that allows designers to tackle the design of very new and challenging products.) What is absolutely fascinating is understanding what happens to the air in flapping wing flight. This is unsteady flow fluid mechanics. You can use computer modelling techniques (CFD) until the cows come home but you ain't gonna really know what goes on with unsteady flow in flapping flight unless you build a working ornithopter and experiment with it. 

.The bird's wing (and bat?) has a hinge half way along its span at its wrist. The bird uses it for what I believe is transitional flight. Hovering flight has more than one mode of operation. Fully developed flight doesn't use the hinge and the flap angle is small. So here are two extremes between hovering flight and fully developed flight. You can't suddenly change the operation of the the wings from hovering mode to fully developed flight mode because the ornithopter will stall. Transitional flight maintains a component of flapping beyond the wrist by virtue of bending at the wrist. The outer portion of the wing as well as giving high lift due to bending on the up-stroke also gives a propulsive component to accelerate the machine up to speed at which point the bending of the wrist can be turned off. I may be wrong on this but it's what I've worked out just by thinking about things....no computer simulations I'm afraid. I guess one analogy is the Harrier jumpjet nozzles swinging from vertical through 45 degrees to horizontal as the aircraft gathers speed and the wings gradually provide more and more lift due to the forward speed rather than lift generated extremely inefficiently by the nozzles pointing downward. By the way, engineering students, why is more power required for the Harrier jumpjet to lift its mass in hovering flight as compared to generating the same lift in forward flight? Take a look at Newton's 2nd law applied to a fluid being accelerated through an area and then look at the kinetic energy in the fluid. What happens to the kinetic energy in the fluid as you decrease the area but still maintain constant lift? In hovering flight the air is accelerated downward through nozzles with a small area. In "fully developed" forward flight the air is being accelerated downward through an area equal to the plan area of the wings which is much greater than the nozzle area. 

It seems that mechanical engineering is holding back research into ornithopters because how are we going to design a wing similar to that of a bird and bat? Look at all the degrees of freedom such as tail warping and wing warping and wing protraction and retraction and wing centre of pressure shifting forwards and backwards. What about getting the wings to flap at frequencies of 5Hz? The power required to do that is enormous unless..... and if that is not a problem what about the high power actuators for the wings? And what about the weight? The bat and bird robot problem must and will be solved sooner or later. Is it too early because technology is not advanced enough? Don't know till you try. Nobody thought anybody could run faster than a 4-minute mile until Roger Bannister made it happen. 

The Challenges 

The technology involved, the constructional techniques, the materials, the dynamical mechanical system, the electronic, computing and control system, the fluid dynamics of the craft, the power plant, the as yet unknown fundamental rules of flapping wing flight are all highly challenging and fascinating. Has anybody built a retracting/protracting, hovering, flapping wing robot? I'm not talking about reciprocating insect type flight. Bird flight surely is a holy grail. 

The Bat Robot 

A lot of people are working on flapping fliers but they are all small fliers of less than 300mm wingspan which is fine but what about bigger flappers? You can go for an insect flapping wing robot (been done for many years by aeromodellers but I've not seen one hovering) and also the latest miniaturised one by a Californian company, Aerovironment.com or you can go for humming bird stiff flapping wing hovering and fully developed flight (the latter been done for 100 years now and the latest by a North American company, Intercept Technologies, but I've also not seen one hovering and I think the control methods still have a long way to go) or you can go for a feathered bird and even use real feathers (difficult this one, not so much the big feathers but also the down feathers, you should pluck the feathers from a chicken wing to see what I mean, but could use contoured foam to give a similar effect) or you can go for a stretched skin bat flapper. This is the one that looks achievable to me. Too small and it is too difficult a challenge; too big and I would say not possible. So what size then? Well how about 2metres for the wingspan? Sounds big but not so when you see it. The wings need to flap at about 5Hz and at the end of the downstroke to retract and then go into an upstroke and then protract and then go into the downstroke thus creating lift and then repeat. Why 5Hz? Well we have some beautiful fliers here on our picturesque NTU campus. They are birds of some 4 to 5 feet of wingspan , I think they are Egrets. They are absolutely beautiful fliers, a sheer delight to watch, an example of incredible flying machines in a slim and neat package. I estimate just by looking and guessing that when they take off and land steeply that they are flapping at about 5Hz. Another interesting thing is that when the wings flap down, the body moves up and vice versa, (recall Newton's third law that "to every action there is an equal and opposite reaction"). I estimate that the ratio is about 10:1 of wing c.ofg.vertical amplitude to body reactionary vertical amplitude which implies that the total wing mass is about 10% of the body mass. This already is a significant design criterion for the design of a flapping flying machine. 

The Bat Robot so far 

We already have built one wing of 1metre in span (this will give an ornithopter of 2metre wingspan). The wing protracts and retracts very close to that of a real bat. The bat has 4 fingers and a thumb like us. The thumb is a short claw and the first two fingers are fused together to produce an up-and-over leading edge aerofoil for the outer half wingspan. The elbow joint is bent down in the plane of the wing so as similarly to produce an up-and-over leading edge for the inner half wingspan. The angles of the fused two fingers and the remaining two fingers of the bat wing linearly protract and retract, with proportional rates, as the wing opens and closes respectively.These features have amazed me because I didn't see them until faced with the problem of constructing a real mechanical wing. It was then that I started asking questions and looking at the detail of these creatures. So here lies a case in point which is that you only really understand a problem when you start to build real hardware. When it comes to friction I have been uncompromising. You know we can run and walk for miles without our joints overheating which implies that the cartilage in our joints must have very low friction. So I have done the same. At every joint that are two ball bearing races. I foresee the whole robot will have about 100 ball bearing races. No they are not too heavy at about 2gm each. A not insignificant weight but due to the flapping wing motion I think friction must be designed to be very low. By the way no CAD package has been used. Why people are obsessed with thinking you can't design without CAD I just don't know. Ignorance of the fundamental design process I guess. Just remember that the ultimate CAD package is the computer on top of your shoulders. Artificial CAD packages cannot replace that one. The brain, coupled with paper and pencil, is your visualiser and your designer, and the manual lathe and milling machine are your rapid prototyping machines. CAD here is deemed to mean computer-aided-design but when it comes to CADraughting then we use these packages such as AutoCad. Whereas I use graph paper and pencil to design components, the workshop technicians would use AutoCad to extract certain dimensions from my drawings that are required in manufacturing the component. 

We are now constructing a collapsible fixed wing, non-flapping, 2m wingspan drone as an interim or off-shoot product. It's a bit like James Bond with a walking stick that, with a quick flick of the wrist, transforms itself into a small spy aeroplane. Later we will get the wings to flap. However, right now the 2m wingspan non-flapping spydrone will use wing and tail warping for its flight control and an electric motor driven propeller for its propulsion. 

The flapping wings of the bat robot need to be able to give control such that the bat can rotate about a vertical axis (yaw), a horizontal athwartships axis (pitch) , and a horizontal longitudinal axis (roll) whilst in hovering flight. The wings need to shift their centres of pressure relative to the centre of gravity forward and backwards and left and right so as to give the pitch and roll motion which in turn will lead to forward/ backward motion and left/right motion. Vertical motion will be by increasing or decreasing the flapping rate. Hence you get 4dof hovering motion. It would be spell-binding to watch a 2m wingspan bat robot hovering in front of your very eyes and jaw dropping to see it manoeuvre up/down, left/right, forward/backward and change heading. Later still the bat robot needs to land in a tree and take off from that tree. Wow we take for granted a bird landing on a compliant bough of a tree. Wouldn't it be something else for a robot to do that. As it waits in the tree it can power down and carry out surveillance and even recharge its batteries if there is any sun light. 

Fred: "Ahem! wake up you silly sod, Alex, stop dreaming and get going". 

Alex: "Wha...wha...whas goin' on ?" 

Fred: "didn't you know you were waving your arms around in your sleep and muttering something about quinny...quinte... .........can't even say it quintessenunreal or something... wassit mean?" 

Alex: "was I saying that? You mean quintessential. I dunno....top-of-the-range, the fifth element after earth, fire, water, and air, the ultimate desire, purest element or something like that. It was just a dream anyway forget it just a dream" 

Fred: "yea yea, you and your dreams. Just remember you do that too much in your sleep an' you will do yourself or someone else an injury" 

Alex: "it's just that I got this urge to make things... " 

Fred: "have you been to see the doctor about this complaint? 

Alex: "yes, but he said it is an incurable disease" 

to be continued