Electromagnetic Actuation for a Dragonfly Inspired Flapping-Wing Micro Aerial Vehicle by Allen Chee A thesis submitted in conformity with the requirements for the degree of Master of Applied Science Graduate Department of the Institute for Aerospace Studies University of Toronto ' Copyright 2017 by Allen Chee Abstract Electromagnetic Actuation for a Dragonfly Inspired Flapping-Wing Micro Aerial Vehicle Allen Chee Master of Applied Science Graduate Department of the Institute for Aerospace Studies University of Toronto 2017 Insect-scale microaerial vehicles is an area within microaerial vehicles which has seen recent growth due to new understandings of insect flight and the availability of new actuation tech- nologies. Prominent flapping wing MAVs were surveyed and relevant observations taken to help guide the project. Alternative actuation technologies for the UTIAS Robotic Drag- onfly project were assessed and an electromagnetic actuator was selected. A new design incorporating this actuator was fabricated was fabricated and tested. The platform fea- tures a sub-gram at-scale prototype with independently driven wings, a mass of 222 mg and a wingspan of 75 mm. Experiments demonstrated that the prototype was capable of generating up to 1.34 mN of lift. ii Contents 1 Introduction 1 1.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.1.1 Mimicking Nature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 1.1.2 Robotic Biomimicry . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 1.2 Flapping-Wing Legacy at UTIAS . . . . . . . . . . . . . . . . . . . . . . . . 4 1.3 UTIAS Robotic Dragonfly . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 1.3.1 Contributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 2 Literature Review 7 2.1 Dragonflies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 2.1.1 Evolution of Dragonflies . . . . . . . . . . . . . . . . . . . . . . . . . 7 2.1.2 Anisoptera Morphology . . . . . . . . . . . . . . . . . . . . . . . . . . 9 2.1.3 Flight Kinematics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 2.1.4 Powering Dragonfly Flight . . . . . . . . . . . . . . . . . . . . . . . . 13 2.1.5 Sympetrum sanguineum . . . . . . . . . . . . . . . . . . . . . . . . . 15 2.2 A Different Flight Regime . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 2.2.1 Dragonfly Simulations . . . . . . . . . . . . . . . . . . . . . . . . . . 17 2.2.2 Scaled Flapping Experiments . . . . . . . . . . . . . . . . . . . . . . 18 2.2.3 Flight Mechanisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 2.3 MAVs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 2.3.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 2.3.2 Application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 2.3.3 Other MAV Projects . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 3 Project Background 41 3.1 Original Goal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 3.1.1 Idealised Dragonfly . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 3.2 Prototype Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 3.2.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 3.2.2 Wings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 3.2.3 Piezoelectric Actuators . . . . . . . . . . . . . . . . . . . . . . . . . . 46 3.2.4 Transmission . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47 3.2.5 Frame . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47 3.2.6 Power . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47 3.2.7 Summary of 2P# Platform . . . . . . . . . . . . . . . . . . . . . . . . 48 3.3 Modelling, Experiments and Discussion . . . . . . . . . . . . . . . . . . . . . 48 iii 3.3.1 Modelling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 3.3.2 Experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 3.3.3 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 4 UTIAS Robotic Dragonfly and the Search for a New Actuator 51 4.1 Relaxation of Constraints . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 4.2 A New Actuator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 4.2.1 Commonly Used Actuators . . . . . . . . . . . . . . . . . . . . . . . . 52 4.2.2 A New Search . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 4.3 Electromagnetic Actuation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 4.3.1 Actuator Characterization . . . . . . . . . . . . . . . . . . . . . . . . 59 4.3.2 Linear Actuation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 4.3.3 Rotational Actuation . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 4.4 Prototype Design & Fabrication . . . . . . . . . . . . . . . . . . . . . . . . . 64 4.4.1 1EM Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 4.4.2 2EM Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68 4.4.3 Simulation Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72 4.5 Lift Measurement Apparatus . . . . . . . . . . . . . . . . . . . . . . . . . . . 74 4.5.1 Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75 4.5.2 Static Calibration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76 4.5.3 Power Supply . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 4.6 Summary of 2EM Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 4.7 Lift Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78 5 Conclusion 83 5.1 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83 5.2 Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84 5.2.1 Simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84 5.2.2 Fabrication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84 5.2.3 Increased Power Density . . . . . . . . . . . . . . . . . . . . . . . . . 84 5.2.4 Imaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 5.3 Contributions and Closing Remarks . . . . . . . . . . . . . . . . . . . . . . . 85 Bibliography 86 iv List of Tables 2.1 Average physical parameters of Sympetrum Sanguineumfrom Wakeling and Ellington [116, 118] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 2.2 PerformanceparametersofSympetrumSanguineum fromWakelingandElling- ton [117][118] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 2.3 Phyiscal parameters of the MFI platform . . . . . . . . . . . . . . . . . . . . 24 2.4 Consecutive DelFly platform iterations [79] . . . . . . . . . . . . . . . . . . . 27 2.5 Comparison of MAV platforms developed at Cornell University [3] . . . . . . 28 2.6 Comparison of MAV platforms developed at Carnegie Mellon University[3] . 30 2.7 Comparison of MAV platforms developed at the University of Deleware and Purdue University . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 2.8 Comparison of MAV platforms developed at Harvard University [6] . . . . . 34 2.9 Comparison of MAV platforms developed at Shanghai Jiao Tong University [148, 149] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 3.1 Body parameters of the Idealised Dragonfly . . . . . . . . . . . . . . . . . . 41 3.2 Physical parameters of the Idealised Dragonfly’s wings . . . . . . . . . . . . 42 3.3 Performance parameters of the Idealised Dragonfly . . . . . . . . . . . . . . 42 3.4 Artificial wing properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 3.5 Body parameters of the Idealised Dragonfly . . . . . . . . . . . . . . . . . . 49 4.1 Body parameters of the Modified Dragonfly . . . . . . . . . . . . . . . . . . 51 4.2 Physical parameters of the Modified Dragonfly’s wings . . . . . . . . . . . . 52 4.3 Performance parameters of the Modified Dragonfly . . . . . . . . . . . . . . 52 4.4 Comparison of sub-gram actuation technologies based on the work of Karpel- son and Bell et al. [71, 29] . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 4.5 Physical parameters of electromagnetic actuator . . . . . . . . . . . . . . . . 59 4.6 Specifications of SMD S215 load cell . . . . . . . . . . . . . . . . . . . . . . 75 4.7 Specifications of the DAQ (MCC USB-1608G) . . . . . . . . . . . . . . . . . 76 4.8 Specifications of the DAQ (MCC USB-1608G) . . . . . . . . . . . . . . . . . 76 4.9 List of Precision masses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76 4.10 Body parameters of the Modified Dragonfly . . . . . . . . . . . . . . . . . . 78 v List of Figures 1.1 Previous examples of flapping-wing projects at UTIAS . . . . . . . . . . . . 4 2.1 Examples of the order Meganisoptera dating to the Carboniferous period . . 8 2.2 Comparison of dragonflies and damselflies . . . . . . . . . . . . . . . . . . . 10 2.3 Actual dragonfly kinematics [123] . . . . . . . . . . . . . . . . . . . . . . . . 12 2.4 Comparison between normal hovering and inclined stroke plane during insect hovering [123] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 2.5 Sympetrum Sanguineum [24] . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 2.6 Schematics showing delayed stall and wake capture during normal hovering [4] 20 2.7 Schematic showing the clap-and-fling [39] . . . . . . . . . . . . . . . . . . . . 21 2.8 MFI platform and custom developed sensors . . . . . . . . . . . . . . . . . . 25 2.9 Consecutive iterations of the DelFly project [79] . . . . . . . . . . . . . . . . 27 2.10 Various MAV platforms developed at Cornell University [89, 115] . . . . . . 28 2.11 Various MAV platforms developed at Cornell University [3] . . . . . . . . . . 30 2.12 Dragonfly-based MAV developed at University of Delaware [48] . . . . . . . . 31 2.13 Various MAVs developed at Purdue University[70, 92] . . . . . . . . . . . . . 32 2.14 Various MAVs developed at Harvard University[138, 6] . . . . . . . . . . . . 33 2.15 Harvard Microrobotics lab’s ”pop-up” assembly from monolithic laminate [103] 35 2.16 Schematic of differential mechanism allowing active control of the mean wing hinge position of the Harvard RoboBee [112] . . . . . . . . . . . . . . . . . . 37 2.17 ElectromagneticallyactuatedMAVplatformdevelopedatShanghaiJiaoTong University [148] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 2.18 Piezoelectric actuated MAV platform developed at Shanghai Jiao Tong Uni- versity [149] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 3.1 Artistic rendition of the ideal UTIAS Robotic Dragonfly . . . . . . . . . . . 42 3.2 Comparison of single and tandem wing pair platforms . . . . . . . . . . . . 44 3.3 Comparison of single and tandem wing pair platforms . . . . . . . . . . . . 46 4.1 Map of force versus displacement for MEMS and macro actuators by Bell et al. [29] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 4.2 Map of frequency versus displacement for MEMS and macro actuators by Bell et al. [29] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55 4.3 Schematic of a parallel plate electrostatic actuator . . . . . . . . . . . . . . . 56 4.4 Schematic of a dielectric elastomer actuator [5] . . . . . . . . . . . . . . . . . 57 4.5 Electromagnetic actuator components . . . . . . . . . . . . . . . . . . . . . . 59 4.6 Magnetic field results of actuator characterization . . . . . . . . . . . . . . . 61 vi 4.7 Linear electromagnetic force due to coil [17] . . . . . . . . . . . . . . . . . . 62 4.8 Ideal magnet position for linear actuation [148] . . . . . . . . . . . . . . . . 63 4.9 Magnetic field along axis of coil[83] . . . . . . . . . . . . . . . . . . . . . . . 64 4.10 Transmission mechanism based on four-bar link mechanics . . . . . . . . . . 66 4.11 Rendition of fabricated subassembly . . . . . . . . . . . . . . . . . . . . . . 66 4.12 Schematic of transmission design featuring an off centre rotation axis [109] . 69 4.13 Schematic of transmission design featuring two parallel joints inside the coil . 70 4.14 Schematic of transmission design featuring fixed beam bending mode . . . . 70 4.15 Schematic of transmission design featuring external joints . . . . . . . . . . . 71 4.16 Orientation of carbon fibre layers . . . . . . . . . . . . . . . . . . . . . . . . 72 4.17 Static calibration of the lift measurement apparatus . . . . . . . . . . . . . . 77 4.18 Experimental results for 2EM14 while driven at 18 Hz . . . . . . . . . . . . . 80 4.19 Experimental results for 2EM15 while driven at 18 Hz . . . . . . . . . . . . . 81 4.20 Experimental results for 2EM18 while driven at 20 Hz . . . . . . . . . . . . . 82 4.21 Experimental results for 2EM19 while driven at 23 Hz . . . . . . . . . . . . . 82 vii Chapter 1 Introduction Mankind’s long history has many accounts of aerial vehicles and devices from ancient tales of Icarus to the development of kites, hot air balloons, and rotor wings. While many of these inventions are fascinating, many of them were unmanned toys and controlled falling. The origins of what we tend to regard as the development of true flight able to carry passengers has much of its origins in a biomimetic flapping wing approach. Through the years there have been attempts at designing flapping wing contraptions using all manner of design. One famous design is Leonardo da Vinci’s flying machines using ropes and pulleys controlled by the pilot to actuate the device [8]. Otto Lillenthal’s kleiner Schlagflu¨gelapparat design strapped wings to the pilots own arms [10]. Robert Hooke even worked on a design using springs as artificial muscles upon the realization that a human muscle power was insufficient for flapping flight [68]. Finally after centuries of different designs, attempts, and unfortu- nately a couple deaths, two brothers, Orville and Wilbur Wright succeeded where others had failed and were the first to achieve a sustained, powered, heavier-than-air, and manned flight [13]. On December 17, 1903, Wilburt Wright piloted the Wright Flyer for 59 s over a distance of 892 ft. Their success has been attributed to the development of aerodynamics, structure, power, and control. Even many years after the Wright brothers’ famous flight and the explosion of air travel, insect flight was poorly understood and thought of as impossible to mimic. Some believe we are approaching a similar culmination of these four areas in regard to recreation of insect flight. Since the days of the Wright brothers there are now high-speed cameras available, many researchers are finally have the technology they require in the pursuit of understanding insect flight. 1.1 Motivation Other than the technological and academic challenge, the development of MAVs has many potential applications. Applications of MAVs could range from search and rescue operations to drone surveillance. A controllable or autonomous MAV would gain certain advantages over currently used robotics. These include the ability to fly and cover adverse terrain as well as operate in close quarter urban settings which traditional UAVs are unable to do. Such a device would also enjoy the properties of being small and light weight excelling at discrete surveillance, or even extraterrestrial exploration. MAVs have the potential to offer advantages over rover-type vehicles and current un- 1 Chapter 1. Introduction 2 manned aerial vehicles (UAVs). In a search and rescue situation, hazardous obstacles and general clutter around the search area may be present that would greatly reduce a traditional rover’s ability to manoeuver but not MAVs [139, 145]. In the field of crowd monitoring, some police departments often use solitary UAVs. These units can be quite expensive costing over $100,000. In the event of damage or malfunction the financial cost is very large along with the potential for complete failure of the missions[138]. Alternatively, a swarm of small MAVs could cover more area, and be more cost effective. If a small portion of the units were lost, the remaining units could continue the mission and the MAVs being so small would not pose any hazardous threat to civilians on the ground. 1.1.1 Mimicking Nature Much of the early inspiration for flight came from birds, a relatively large creature which takes advantage of steady lift mechanisms (time independent flow). Insects, on the other hand have evolved to use unsteady lift mechanisms (time dependent flow) which can only be exploited at their small size. Owing to the wings being small, passive, rigid structures actuated from muscles within the thorax, insects are able to flap their wings at much higher frequencies than than birds can [38]. If one were to attempt to create a traditional fixed- wing aircraft that was the size of an insect, the vehicle would have to travel at speeds far beyond what is currently possible [37]. One obvious solution to this challenge is to attempt to exploit the unsteady lift mechanisms used by insects with flapping wings. Insect wings have a unique place in the arena of flying animals. All other flying animals have evolved wings that are actually modified legs. In the case of birds and bats, their wings have the ability to change shape as well as fold up to be more compact while giving up a set of limbs. The insects on the other hand form wings from modified portions of their exoskeletons. The wings are fixed and nonliving much like human hair and fingernails are nonliving [22]. Insects were the first to develop flight and are quite different from other flying creatures. The earliest insects developed flight over 350 million years ago, and have a striking resemblance to currently extant dragonflies [54]. Both these ancient and modern insects the flight muscles were simple and were made up of two pairs. Modern dragonflies actually seem to be nearly identical in physical configuration [38]. Some insects like flies and bees, on the other hand, have evolved more recently and make use of more complex wing strokes such as the clap-and-fling or stroke-plan deviation, which is rarely observed in dragonflies. Despite potentially appearing to be less evolved than other insects, dragonflies are still considered to be some of the most manoeuvrable and predatory flying insects in existence. It is interesting that many insects maximize efficiency by flapping at resonance based on the structure of the thorax. It seems insects have found a way to contrast starkly the many first year engineering lectures where resonance is often associated with the epic Chapter 1. Introduction 3 failure with many projects. Similar to these types of cases, insects can create large flapping amplitudes from minimal energy cost[31]. In some cases, attempting to mimic nature may help understanding nature and lend weight to scientific hypotheses. Currently, the origins and development is still strongly contested. One theory that has been suggested is that running animals grew protrusions out of their thoraxes to increase inertial stability [38]. These protrusions could have allowed insects with this trait to move with more stability and have an advantage in escaping from predators, or accessing food and mates. Over time a membrane might develop turning what were once protrusions into wings allowing for insects to run and glide and eventually flap and take-off. In 2011, a team led by Ronald Fearing experimented with robotics that tested this idea. The team had developed a bipedal running robots that would lose stability and fall over when running at high speeds [1]. Consecutive modifications of the robot added spars, fixed wings and flapping wings which each gave the robot increased stability and improved its ability to run up steeper and steeper inclined planes [86]. The various hypotheses of insect wing origins only have limited fossil evidence supporting their claims, now this one theory has some form experimental robotic validation [85]. This kind of work demonstrates how ideas from one field can affect another and perhaps hint towards greater productivity if the fields of robotics and evolutionary biology collaborated in the future. 1.1.2 Robotic Biomimicry While some have loosely described MAVs as being less than 30 cm and less than 100 g, the term MAV is most commonly applied to flying robots being less than 20 cm in largest di- mension [135]. This is description is quite vague, but the simple size constraint has naturally led to the majority of projects towards biologically inspired flapping-wing MAVs. Over the years there has been quite a few flapping-wing MAV projects. Only a few attempt to be truly biomimetic, while the remaining tend to be loosely mimicking insect flight. One MAV that fits into the latter category is the DelFly Micro from Delft University of Technology. The MAV features to pairs of flapping wings but also a rudder and elevator. The Delfly has gained some success in flapping flight but does not actively attempt to mimic any of the characteristics found in nature such as wing design, kinematics, mass, or control surfaces. The earliest attempt to make a truly biomimetic MAV was in the 1990s by the Univer- sity of California, Berkeley called the Micromechanical Flying Insect (MFI)[6]. The project aimed at being of a similar size and weight of real insects. It could control its wings with two degrees of freedom but was unable to achieve lift-off. The MFI project was instrumental in defining and developing the field of flapping wing MAVs. In fact, many of the researchers on the project are now leading researchers in the field at other universities.