A flapping feathered wing-powered aerial vehicle
11 School of computer, electronic & information engineering, Guangxi University; Nanning, GuangXi, China. Bee-eater Technology Inc.; Nanning, GuangXi, China. School of electrical engineering, Guangxi University; Nanning, GuangXi, China.*Corresponding author. Email:[email protected]
A flapping feathered wing-powered aerial vehicle
Zhenhong Zhang , Aiqiu Wei , Yu Cai * An aerial vehicle powered by flapping feathered wings was designed, developed andfabricated. Different from legacy flapping-wing aerial vehicles with membrane wings, thenew design uses authentic bird feathers to fabricate wings. In field tests, a radio-controlledelectric-powered aerial vehicle with flapping feathered wings successfully took off, flew up to63.88 s and landed safely. It was found that flapping feathered wings can generate sufficientthrust and lift to make a man-made aerial vehicle accomplish takeoff, sustainable flight anda safe landing.
INTRODUCTION
In 1874, Alphonse Penaud (1) developed the first rubber-band-powered flyable flapping-wingaerial vehicle. Later, a commercial product, named Tim Bird (2) , powered by rubber bands, onwhich many of the following ornithopter designs are based, was released to the market. In recentyears, many bionic flapping-wing robotic platforms inspired by flying creatures have beendeveloped (3) . Among them, Robofly (4,5) and Delfly (3,6) are well-known examples ofinsect-level flapping-wing aerial vehicles. Birds, the most popular flying animals, attract moreattention than other creatures (7-9) . Many successful flyable bird-like flapping-wing aerialvehicles have been invented; for example, Slow Hawk (10) uses a radio-controlled (RC) motorand transmission system to control the synchronously flapping of wings, which generateaerodynamic power to make the robotic bird fly. In subsequent research, many ornithopters withstructures similar to that of Slow Hawk were developed, among which Phoenix (11) is a typicalexample. Another typical ornithopter is Robo Raven (12) . Unlike Slow Hawk, Robo Raven usestwo independent RC servo motors to independently control the flapping of its wings. For SlowHawk, the flapping of the two wings is synchronized, which means the flapping is in-phase,in-rate and interdependent (10) . The wings of either Slow Hawk or Robo Raven are fabricatedwith carbon fiber skeletons covered with a plastic foam membrane or thin plastic film (13,14) . In2011, an ornithopter named Smartbird was developed by Festo (15-17) and successfully flew.Smartbird has a novel hinged-wing design. Each wing consists of two sections: the inner wing andouter wing. The inner wing is connected to the outer wing by hinges (15) . For each wing, there aretwo or three meticulously designed four-bar linkage systems; the power is transmitted through thetorso to the two wing sections, making each wing section flap with different kinematics. Thedifferent movements of the inner and outer wings makes Smartbird extremely authentic. Althoughts wings look like 2-degree-of-freedom (DOF) robotic bird wings, the movements of the innerwing and outer wing are not independently driven, and their movements are interrelated so that theoverall DOF is less than 2 (15) . Either the inner wing or outer wing of Smartbird is made ofmembrane-like materials, including extruded polyurethane foam (17) , which is different from abird’s feathered-wing structure. In 2020, Lentink’s team developed a flyable aerial vehicle calledPigeonBot (19) , whose wings are made with real bird feathers. This demonstrated that theasymmetric morphing of wings and different feather positions can control the flight direction.Unlike birds powered by flapping wings, PigeonBot is powered by a propeller. Creating a flyableflapping-wing robot with feathered and deformable wings has always been an open question (19,20) . This research implemented the idea of implanting real bird feathers on a hinged-wingrobot and utilizing flapping feathers to generate aerodynamic power. The robot powered byflapping feathered wings successfully took off and flew for more than 60 s and then ultimatelyland safely. Field tests proved that carefully designed flapping feathers can generate significantaerodynamic power to drive man-made aerial vehicles.
WINDRIDER TEST PLATFORM
To ensure successful takeoff and sustainable flight, the structure of an aerial vehicle has to belightweight and strong. The authors compared various materials and finally selected a 1.5 mmthick carbon fiber sheet to build torso frames and wing ribs. The backbone of the torso is made ofa 6 mm carbon fiber rod, and the backbone of the wing is made of a 4 mm carbon fiber rod. Partsof the load-bearing points and joints are made with 6063 aluminum alloy. The regular parts weremade with a 3D printer. The printing material was polylactic acid (PLA). More details are shownin Fig. 1.The designed robot is named WindRider, and its design is based on that of Smartbird. Theauthors designed a carbon fiber frame for the main transverse frame of the robot. The motor,battery, transmission gear and four-bar linkages are housed on the main transverse frame and torsobackbone. The running motor drives the gearbox with a 1:48 reduction ratio. After speed reduction,the output of the gearbox drives the wings up and down via a four-bar linkage system (Fig. 1E).Essentially, there are two four-bar linkages in the transmission system, one driving the inner wingsection and the other driving the outer wing section. The second four-bar linkage couples upon thefirst four-bar linkage; therefore, they are interrelated. The inner wing generates lift, and the outerwing generates thrust (15) . A 3 mm thick foaming rate of 45 expanded polypropylene (EPP) filmattached to the tail frame by foam glue (18) was used as the vertical stabilizer. The tail of the aerialvehicle is controlled by RC transceiver systems WFT07 and WFR07S (24) to accomplish the pitchnd yaw movements of the robot (Fig. 1D). When the tail rotates around the elevator axis (Fig.1B), a pitch moment is generated (Fig. 1D) to adjust the pitch angle of the robot. When the tailrotates around the rudder axis (Fig. 1B), a yaw moment is generated around the yaw axis of therobot (Fig. 1D), causing the vehicle to turn left or right (17) . According to the field experience ofthe controller of WindRider, who is also one of authors, qualitatively, the yaw controlling torquecan be partly coupled with the roll controlling portion. Because coupling effects are insignificant,it is difficult for robots to make sharp turns or other acrobatic movements. More quantitativeresearch on this phenomenon is required in future studies. To protect the torso frame and reduceaerodynamic drag, a streamline seagull-like EPP outer shell (Fig. 1C) was fabricated. The torsoframes and outer shell are adhered with foam glue. The final version of WindRider’s skeleton isshown in Fig. 1 E to F. ig. 1. WindRider: A bionic flapping-wing aerial vehicle with feathered wings that consists of a torso, a tailand hinged feathered wings. (A). Top and bottom views of WindRider. It has a wingspan of 195 cm, an innerwingspan of 32 cm, an outer wingspan of 65 cm, a length (from nose to tail) of 104 cm, an overall weight of 667.2g (including a 3-cell Li-po battery with a capacity of 350 mAh and rated voltage of 11.7 V), and a takeoff flappingfrequency of ~ 4 Hz. (B). The tail has a vertical stabilizer, and the vertical stabilizer is attached to the tail frame tomake vehicle left or right turn. (C). The seagull-like torso is made of EPP and carved by CNCs (computernumerical control). (D). The 3 axes of the robot are roll, pitch, and yaw. (E). The front view of the skeleton and thetransverse frame of the torso, cranks, and wing ribs are made of carbon fiber sheets carved by a CNC machine. Thegearbox is made of 6063 aluminum alloy. (F). The top view of the skeleton. The tail and the torso are connected bya 3D-printed plastic part. The maximum continuous power of the motor (Sunnysky-2206,1900 kV) is 150 W. Twoservos (EMAX ES09MD with a torque of 2.3/2.6 kg.cm and a speed of 0.10/0.08 s/60°) are used to control the tailduring yaw and pitch movements. The wing ribs are airfoil-like shapes supported by 3 mm and 4 mm carbon fiberrods. The inner wing ribs and outer wing ribs have different shapes, as shown in Fig. 1F.
FIELD TESTS AND PREPARATIONS
The authors designed and developed a reusable wing with a carbon fiber skeleton and authenticbird feathers for the covering material. Before making this complicated feathered wing, theauthors ordered a large number of feathered fans from Taobao.com (23) and disassembled thefeathers as raw materials. After many studies on the feather arrangement of bird wings (21,22) andafter trying different configurations of bird feathers on the wing skeletons, 3 different types ofeathers (Fig. 2C) were chosen. Each single piece of feather was carefully preened and cleanedbefore it was attached to the wing skeleton (Fig. 2A).To verify that the power generated by the feathered wings was sufficient to make the robot fly,the team designed 3 field test steps. The wings used in the field were transformed from legacymembrane wings to feathered wings step by step, and the impacts arising from each subtlemodification on the wings were reflected in the results of each step in the field test.The robot in step 1 is armed with an inner membrane wing and an outer membrane wing. Instep 2, the outer wing is replaced by a feathered wing, and in step 3, both wings are replaced byfeathered wings. The wings in each step are fabricated as follows:
Feathered wing design and fabricationStep 1: Both the inner and outer wings are membrane wings.
The wing skeletons are coveredwith flexible and light EPP films (Fig. 2F. 1). To ensure that the inner wing generated sufficient liftwhile minimizing the weight, the team chose a 3 mm thick, high foaming rate (45 times) thin EPPfilm to cover the inner wing. Because the outer wing generates thrust and requires more flexibilityduring flapping movements, a 3 mm thick EPP film with a low foaming rate (30 times) is chosen.A low foaming rate results in more flexibility and more weight.
Step 2: The inner wing is a membrane wing, and the outer wing is a feathered wing.
Thewing skeletons of both wing sections and the membrane covering materials on the inner wingremain unchanged. The supporting beams of the outer wing skeleton are 4 mm hollow carbonfiber rods connected with large outer wing ribs; it is difficult to cover the entire surface withfeathers. To solve this issue, part of the outer wing covering the membrane is reserved, as shownin Fig. 2E, as a supporting platform. The wing rib of the outer wing is reserved and unchanged.The reserved membrane (Region R in Fig. 2E) imitates the muscles on a bird’s wing tips andfacilitates feather attachment on the wing skeleton. It avoids the intentional overlap of the flappingregions of the primary flying feathers. Therefore, the reserved membrane itself does not contributesignificant aerodynamic force during flapping movements. Three kinds of selected feathers (Fig.2C) are glued to the R region of the outer wing to imitate the feather arrangement of a real bird.The long brown feathers serve as the primary flying feathers. The short black feathers resemblethe covering feathers of a real bird and are attached to the R region of the outer wing. The finalpicture of the feathered outer wing is shown in Fig. 2F.2.
Step 3: Both wings are feathered wings.
After the field test that used WindRider with featheredouter wings was performed, the inner wing was ready to be feathered. As in step 2, the membranein the G region is reserved to facilitate feather attachment on the skeleton (Fig. 2E). The longbrown feathers (Fig. 2C) were glued to the lower half of the G region to serve as secondary flyingeathers, while the short black feathers were glued to the upper half of the G region as a secondarycovering feathers. The final picture of the inner feathered wing is shown in Fig. 2F. 3.Fabrication of the feathered wing proved to be very challenging in the sense that the amount ofglue on each feather had to be carefully controlled to avoid adding too much weight. The wetregion, where a feather is wetted by glue, of every feather had to be precisely controlled to preventthe feathers from losing flexibility due to their subtle feather textures sticking together and turninginto rigid boards. WindRider with feathered wings is shown in Movie 1.
Fig. 2. Raw feathers, feathered wings and the iterative fabrication process. (A). Final version of the featheredwings. (B). The feathers in the picture are colored. The original colors are shown in (A) and (C). (C). The rawfeathers are categorized into 3 types in terms of the different feather lengths. From left to right: little coverts, largecoverts, and primary flying feathers (or secondary flying feathers). (D). The three types of feathers used in (B).Each type of feather is colored in the picture to map the feather to its covering region. (E). The original membranewing (left picture). The reserved foam film to support feather attachment (right picture). The reserved region in theright picture is colored. (F). The feathered wing fabrication requires 3 iterations, and each iteration gives rise to anew generation of wings. From the previous generation to the new generation requires successful field tests torove the aerodynamic feasibility. The outcomes of the iteration steps are shown in the pictures: (1) foam filmcovering the inner and outer wing skeletons; (2) the inner wing with a membrane structure and the outer wingcovered in real bird feathers; and (3) both the inner and outer wings covered in authentic bird feathers. The wingsshown in pictures (2) and (3) share the same skeletons as the wings in picture (1). The feathers in picture (2)appear to be darker than those in picture (3) because of the light source and background in the picture. They areessentially the same color.
Movie 1. Three gestures of the feathered wings of WindRider . Low-frequency flapping (1 Hz) demonstrates theflapping kinematics of WindRider in flight. The face of the person holding the aircraft is intentionally blurred.
RESULTS
In step 1 of the test, both the inner and outer wings of WindRider were covered with a foam filmmembrane; WindRider accomplished a successful flight up to 150 s (Movie 2). In step 2 of the test,the outer wings were converted to feathered wings, and the inner wings remained unchanged. Themembrane inner wings generated gliding lift, and the outer wings covered with authentic birdfeathers provided flapping thrust. On average, the successful flight lasted for 90 s (Movie 3). Instep 3 of the test, all wings were covered with authentic bird feathers. The flapping-wing aerialvehicle successfully took off and sustainably flew for up to 63.88 s until it finally landed safely.(Movie 4) proved that the power generated by flapping wings covered in authentic bird feathers issufficient to make a man-made robot fly.
Movie 2. Field test of WindRider with membrane wings.
Both the inner and outer wings of the robot arecovered with a foam film membrane. Its weight is of 588.0 g. The flight duration is 168.88 s. The attitude iscontrolled by the tail, and landing occurs via unpowered gliding. ovie 3. In step 2 of the field test, the inner wing skeleton is covered with foam film and the outer wing iscovered with feathers.
The weight of the whole machine is 630.8 g. The takeoff flapping frequency isapproximately 4 Hz. WindRider takes off from an open field and hovers and rises by flapping its wings. Its flight ismanipulated by an RC flyer via WFT07. Landing is achieved by reducing the flapping rate until the abdomentouches the ground. The flight duration (from takeoff to landing) is 90.04 s.
Movie 4. WindRider in step 3 of the field test; the wings are all covered in authentic bird feathers.
Theweight is 667.2 g. The takeoff flapping frequency is ~ 4 Hz. The flight duration (from takeoff to landing) is 63.88 s.The RC flyer controls the aircraft while it hovers in the air; finally, the wing flapping frequency is reduced untilWindRider lands safely.
CONCLUSIONS AND FUTURE WORK
This paper proposes a flapping-wing aerial vehicle whose wings are covered in authentic birdfeathers. Different from legacy flapping-wing aerial vehicles, whose wings are membranestructures, the authors replaced the wing membranes with authentic bird feathers. With flappingfeathered wings, the robot obtained sufficient lift and thrust in order to take off, fly and land safely.The flight lasted up to 63.88 s. The research demonstrates that flapping wings with carefullyarranged bird feathers can generate sufficient aerodynamic force to make a man-made robot fly. Inthe future, the authors will cover the whole torso and tail with bird feathers to make the robotore similar to a real bird, regardless of the outlook perspective or aerodynamic perspective.
REFERENCES AND NOTES
1. J. D. DeLaurier, An ornithopter wing design.
Canadian aeronautics and space journal. ,10-18 (1994).2. J. D. DeLaurier, J. M. Harris, A study of mechanical flapping-wing flight. The AeronauticalJournal . , 277-286 (1993).3. M. Karásek, F. T. Muijres, C. De Wagter, B. D. Remes, G. C. de Croon, A tailless aerialrobotic flapper reveals that flies use torque coupling in rapid banked turns. Science ,1089-1094 (2018).4. K. Y. Ma, P. Chirarattananon, S. B. Fuller, R. J. Wood, Controlled flight of a biologicallyinspired, insect-scale robot.
Science , 603-607 (2013).5. R. J. Wood, The first takeoff of a biologically inspired at-scale robotic insect.
IEEEtransactions on robotics. , 341-347 (2008).6. G. C. H. E. De Croon, K. M. E. De Clercq, R. Ruijsink, B. Remes, C. De Wagter, Design,aerodynamics, and vision-based control of the DelFly. International Journal of Micro AirVehicles. , 71-97 (2009).7. C. J. Pennycuick, Modelling the Flying Bird (Elsevier Science, 2008).8. D. Lentink, U. K. Müller, E. J. Stamhuis, R. de Kat, W. van Gestel, L. L. Veldhuis, P.Henningsson, A. Hedenström, J. J. Videler, J. L. van Leeuwen, How swifts control their glideperformance with morphing wings.
Nature , 1082–1085 (2007).9. H. Weimerskirch, C. Bishop, T. Jeanniard-du-Dot, A. Prudor, G. Sachs, Frigate birds trackatmospheric conditions over months-long transoceanic flights.
Science , 74-78 (2016).10. A. Kinkade, Ornithopter, U.S. Patent US20020173217A1 (2002).11. R. Tedrake, Z. Jackowski, R. Cory, J. W. Roberts, W. Hoburg, “Learning to fly like a bird.”In . (2009).12. J. Gerdes, A. Holness, A. Perez-Rosado, L. Roberts, A. Greisinger, E. Barnett, J. Kempny, D.Lingam, C.-H. Yeh, H. A. Bruck, S. K. Gupta, Robo Raven: a flapping-wing air vehicle withhighly compliant and independently controlled wings.
Soft Robotics. , 275-288 (2014).13. M. F. B. Abas, A. S. B. M. Rafie, H. B. Yusoff, K. A. B. Ahmad, Flapping wingmicro-aerial-vehicle: kinematics, membranes, and flapping mechanisms of ornithopter andinsect flight. Chinese Journal of Aeronautics. , 1159-1177 (2016).14. R. L. Harmon, “Aerodynamic modeling of a flapping membrane wing using motion trackingexperiments”, thesis, University of Maryland (2008).15. W. Send, M. Fischer, K. Jebens, R. Mugrauer, A. Nagarathinam, F. Scharstein, “Artificialhinged-wing bird with active torsion and partially linear kinematics.” In Proceeding of . (2012).16. D. Mackenzie, A flapping of wings. Science.
Science Robotics . , (2020).20. D. D. Chin, L. Y. Matloff, A. K. Stowers, E. R. Tucci, D. Lentink, Inspiration for wing design:How forelimb specialization enables active flight in modern vertebrates. J. R. Soc. Interface , 20170240 (2017).21. T. L. Hieronymus, Flight feather attachment in rock pigeons (Columba livia): Covert feathersand smooth muscle coordinate a morphing wing. J. Anat.
Science367