No description available.

Both bird and fish locomotion are thought to be more efficient than the equivalent man-made vehicles driven by propellers/impellers and jet engines. Through studies that decompose the different kinematic and shape effects of these biological systems, we can understand what leads to their high cruising performance and efficiency. Two major studies were conducted. The first was on the effect of different kinematic parameters of large soaring birds on flight performance and the second was on the effect of caudal fin shape on the performance of thunniform swimmers. For the first study on flight performance, flapping, folding, and twist were the wing motions of interest. The second study on swimming performance observed how caudal fin sweep angle affects propulsion while isolating the effect of this shape difference from aspect ratio and area effects. Low order models were primarily used to conduct the bird flight study, though experimental methods were investigated as well. The thunniform swimming study was conducted through experimentation on a biomimetic system. The flight study found that, under the right circumstances, both wing twist and wing folding have a positive effect on flight performance. However, the impact of wing twist is much larger. To incorporate this wing twist into a robotic system, a new reduced order model that partially accounts for 3D effects was developed and validated. In the future, this model can be used in conjunction with a flight controller to control wing twist. The swimming study found that caudal fin sweep had a significant impact on performance, moderately swept fins showing the greatest improvement. Using an overly large sweep angle led to diminished performance when compared to the moderately swept fins, but still demonstrated improved performance over a non-swept fin. The increased performance of the moderately swept fins was due to how it affected LEV formation and stability. / Ph. D. / Bird flight and fish swimming are thought to be more efficient than drones and submersible vehicles respectively. By conducting studies on the motion of the wing and the shape of the tail fin, we can gain a better understanding of how to produce efficient vehicles that are inspired by fish/birds. Two major studies were conducted. The first study analyzed the wing motion of birds such as seagulls. The three most important wing motions were analyzed using fast computational simulations. Functional flapping aircraft that can be used in future studies were also constructed. The second study analyzed the tail fin shape of tuna, specifically how the swept shape affects propulsion performance. This study was conducted by operating a robotic tuna with interchangeable tails in a water tunnel. The computational studies on wing motion showed that controlling twist of the wing in addition to typical flapping motion could greatly improve performance of a flapping bird-like aerial vehicle. To incorporate this wing twist into a future system, a mathematical model that provided aerodynamic predictions was developed. This model can be used in conjunction with a controller to provide efficient real time control of the wing twist. The experimental swimming study found that fin sweep had a significant impact on performance. Using a moderately swept fin (25-35 degrees) increases thrust production without increased energy expenditure. Fins with greater sweep angles start to yield diminished performance benefits. Using an elliptical area distribution can also lead to increased performance.

Unmanned underwater vehicles (UUVs) have long been in use but increasingly there has been a wave of biomimetic robots taking over the duties and functions of traditional vehicles. A robotic jellyfish, inspired by the species Aurelia aurita was developed and characterized. In addition to the body of the main robotic vehicle, supporting technologies were developed including polymeric artificial muscles, hydrogel-based artificial mesoglea, and an inclinometer inspired by the jellyfish statocyst organ. Through multiple versions, the vehicle was able to attain an order of magnitude increase in proficiency from 0.022 s?? to 0.21 s?? and robustness not found in initial prototypes. A polyvinyl alcohol hydrogel reinforced with ferritin nanoparticles was found to accurately mimic the stress and strain characteristics of natural Aurelia mesoglea while maintaining a high water content similar to the animal. In addition, the optical properties were shown to be controlled by water to DMSO ratio. A five layer PPy-Au-PVDF-Au-PPy actuator stored in 0.5M KCl solution actuated at 4 VDC potential and produced an impressive 90% tip deflection. In addition, the rate of change was extremely high at 50% deflection of initial actuator length per second. The artificial jellyfish statocyst was found to produce the required highly linear voltage divider output. This sensor will provide the vehicle with biomimetic self-awareness of its own body position. Future directions are proposed for the biomimetic robotic jellyfish such as on-board power and computing, multi-material mesoglea with a dermal layer, a MEMS-based statocyst, and polymeric muscles with increased force production and time response. / Master of Science

The design of unmanned underwater vehicle (UUV) was inspired by the form and functionality of Jellyfish. These natural organisms were chosen as bio-inspiration for a multitude of reasons including: efficiency, good room for payload, and a wide range of sizes and morphology. Shape memory alloy (SMA) actuators were selected as the primary source of actuation for the propulsion of the artificial jellyfish node. These actuators offer high power density which enables a compact system size and silent operation which is preferred for surveillance.  SMA wires mimic the form and function of natural muscles; allowing for a wider range of applications than conventional actuators. Commercial SMA wires (100 um in diameter) can exhibit a 4% deformation of the initial actuator length with a blocking stress of over 200 MPa. The deformation of SMA wire is not enough to mimic the bell contraction of jellyfish. In order to resolve this problem, a beam-shape composite actuator using SMA wires as the active component, termed as BISMAC, was designed to provide large curvature. The BISMAC design was inspired by rowing jellyfish bell contraction. Characterization of maximum deformation in underwater conditions was performed for different actuator configurations to analyze the effect of design parameters that include silicone thickness, flexible steel thickness and distance between SMA and flexible steel. A constant cross-section (CC) BISMAC of 16 cm in length was found to achieve deformation with a radius of curvature of 3.5 cm. Under equilibrium conditions, the CC-BISMAC was found to achieve 80% of maximum deformation consuming 7.9 J per cycle driven at 16.2 V/0.98 A and frequency of 0.25 Hz. Using the a developed analytical model, an actuator design was fabricated mimicking the maximum deformation profile of the A. aurita. The optimized AA-BISMAC achieved a maximum curvature of 0.428 1/cm as compared to 0.438 1/cm for the A. aurita with an average squared root error of 0.043 (1/cm), 10.2% of maximum A. aurita curvature.   BISMAC actuators are unidirectional flexible actuators capable of exhibiting high curvature. To extend the application range of these actuators, they were modified to achieve bidirectional deformation. The new bidirectional actuators termed as "BiFlex" actuators had the capability to achieve large deformation in two directions. The FlexLegs consist of six segments which can be actuated individually. Two different sets of legs were constructed to determine the effect of size. The small legs measured 35.8 mm in height and 63.2 mm in width and the large legs were 97.4 mm in height and 165.4 mm in width. The small FlexLegs achieved a maximum deformation of 12 % and 4 % in the x- and y-direction respectively using a power of 0.7 W while producing a maximum force of 0.023 N. They were also able to withstand a load of 1.18 N. The large FlexLegs had a maximum deformation of 57 % and 39 % in the x- and y-direction respectively using a power of 3 W while producing a force of 0.045 N. They were able to withstand a load of 0.25 N. The legs were also able to perform several walking algorithms consisting of stepping, crabbing and yawing. In order to reduce the power consumption and contraction time of SMA wires, a feedback control scheme using wire resistance was developed. The controller required the knowledge of threshold resistance and safe current inputs which were determined experimentally. The overheating effect of SMA wires was analyzed for BioMetal Fiber (BMF) and Flexinol 100 "m diameter wires revealing an increase in resistance as the wires overheated. The controller was first characterized on a SMA wire with bias spring system for a BMF 100 using I_hi=0.5 A and I_low=0.2 A, where hi corresponds to peak current for fast actuation and low corresponds to the safe current which prevents overheating and maintains desired deformation. A contraction of 4.59% was achieved in 0.06 s using the controller and the deformation was maintained for 2 s at low current. The BISMAC actuator was operated using the controller with I_hi=1.1 A and I_low=0.65 A achieving a 67% decrease in contraction time compared to using a constant driving current of I_low=0.2 A and a 60% decrease in energy consumption compared to using constant I_hi=0.5 A while still exceeding the contraction requirements of the Aurelia aurita. Two fundamental parameters at the composition level were associated with the power consumption of SMA: i) martensite to austentite phase transition temperature and ii) thermal hysteresis. Ideally, one would like to reduce both these quantities and for this purpose an equiatomic Ni-Ti alloy was modified with Cu. Replacing nickel with 10 at% copper reduces the thermal hysteresis by 50% or more. For Ni-Ti alloys with nickel content greater than 50 at%, transition temperature decreases linearly at a rate of 100 "C/Ni at%. Given these two power reducing factors, an alloy with composition of Ni40+xTi50-xCu10 was synthesized with x = 0, ±1, ±2, ±3, ±4, ±5. Metal powders were melted in an argon atmosphere using an RF induction furnace to produce ingots. All the synthesized samples were characterized by differential scanning calorimetric (DSC) analysis to reveal martensite to austenite and austenite to martensite transition temperatures during heating and cooling cycles respectively. Scanning electron microscopy (SEM) was conducted to identify the density and microstructure of the fractured samples. The results show the possibility of achieving low power consuming high performance SMAs. Using the BISMAC actuator and feedback control system, a robotic jellyfish called Robojelly that mimics the morphology and kinematics of the Aurelia aurita species was created. A systematic fabrication technique was developed to replicate the essential structural features of A. aurita. Robojelly's body was fabricated from RTV silicone having a total mass of 242 g and bell diameter of 16.4 cm. Robojelly was able to generate enough thrust in static water conditions to propel itself and achieve a proficiency of 0.19 s-1 while the A. aurita achieves a proficiency of around 0.25 s-1. A thrust analysis based on empirical measurements for natural jellyfish was used to compare the performance of the different robotic configurations. The configuration with best performance was a Robojelly with segmented bell and a passive flap structure. Robojelly was found to consume an average power on the order of 17 W with the actuators not having fully reached thermal steady state. A comparative kinematics analysis was conducted between a natural Aurelia aurita and Robojelly. The resistance feedback controller was implemented to tailor the deformation profile of BISMAC actuators embedded in Robojelly. Robojelly's performance was quantified in terms of thrust production and power consumption during vertical swimming experiments. A maximum average instantaneous thrust production of 0.006 N was achieved at a driving current (Ihi) of 1.5 A with 35% duty cycle. Rapid heating of SMA wires was found to reduce power consumption and increase thrust. The bell kinematic analysis revealed resemblance and differences in bell deformation trajectories of the biomimetic and natural jellyfish. The inflexion point of the A. aurita was found to convert an inner bell trajectory into an outer one during contraction which assists the thrust production. A biomimetic robot inspired by Cyanea capillata, termed as "Cyro", was developed to meet the functional demands of underwater surveillance in defense and civilian applications. The design of Cyro required kinematics of large C. capillata which are elusive creatures. Obtaining accurate kinematic data of animals is essential for many biological studies and bio-inspired engineering applications. Many animals such as the C. capillata however, are either too large or too delicate to transport to controlled environments where accurate kinematic data can easily be obtained. Often, in situ recordings are the only means available but are often subject to multi-axis motion and relative magnification changes with time, which lead to large discrepancies in animal kinematics. In Chapter 5, techniques to compensate for magnification and body rotation of animal footage were developed. A background reference point and animal dimensions were used to account for magnification. A linear fit of body points were used to measure body rotation. These techniques help resolve animal kinematics from in situ video footage. The techniques were applied to a large jellyfish, Cyanea capillata, swimming in ocean waters. The bell kinematics were captured by digitizing exumbrella profiles for two full swimming cycles. Magnification was accounted for by tracking a reference point on the ocean floor and by tracking the C. capillata exumbrella arclength in order to have a constant scale through the swimming cycles. A linear fit of the top bell portion was used to find the body angle with respect to the camera coordinate system. Bell margin trajectories over two swimming cycles confirm the accuracy of the correction techniques. The corrected profiles were filtered and interpolated to provide a set of time-dependent points along the bell. The ability to use in situ footage with significant multi-axis motion provides an opportunity to analyze previously impractical footage for gaining a better understanding of large or delicate organisms. The swimming kinematics of the C. capillata were analyzed after extracting the required kinematics from the in situ video. A discrete model of the exumbrella was developed and used to analyze the kinematics. The exumbrella discretization was done using three different methods. The first method consists of analyzing the animal anatomy for structural and mechanical features. The second method consists of analyzing the bell kinematics for areas of highest deformation over time. The third method consists of optimizing node locations that can provide minimal error with comparison to the digitized profiles. Two kinematic models of the C. capillata swimming motion were developed by fitting Fourier series to the discretized segments and angles formed by each segment. The four-segment anatomical kinematic model was used to analyze the bell kinematics of the C. capillata. It was found that the bell does not deform uniformly over time with segments lagging behind others. Hysteresis between contraction and relaxation was also present through most of the exumbrella. The bell margin had the largest hysteresis with an outer path during contraction and inner path during relaxation. The subumbrella volume was approximated based on the exumbrella kinematics and was found to increase during contraction. Cyro was designed to mimic the morphology and swimming mechanism of the natural counterpart. The body of the vehicle consists of a rigid support structure with linear DC motors which actuate eight mechanical arms. The mechanical arms in conjunction with artificial mesoglea create the hydrodynamic force required for propulsion. The full vehicle measures 170 cm in diameter and has a total mass of 76 kg. An analytical model of the mechanical arm kinematics was developed. The analytical and experimental bell kinematics were analyzed and compared to the C. capillata. Cyro reached the water surface untethered and autonomously from a depth of 182 cm in five actuation cycles. It achieved an average velocity of 8.47 cm/s while consuming an average power of 70 W. A thrust stand was developed to calculate the thrust directly from a single bell segment yielding an average thrust of 27.9 N for the whole vehicle. Steady state velocity during Cyro's swimming test was not reached but the measured performance during its last swim cycle resulted in a cost of transport of 10.9 J/kg m and total efficiency of 3%. It was observed that a passive flexible margin or flap, drastically increases the performance of the Robojelly. The effects of flap length and geometry on Robojelly were analyzed using PIV. The flap was defined as the bell section which is located between the flexion point and bell margin. The flexion point was established as the location where the bell undergoes a significant change compliance and therefore in slope. The flap was analyzed in terms of its kinematics and hydrodynamic contribution. An outer trajectory is achieved by the flap margin during contraction while an inner trajectory is achieved during relaxation. The flap kinematics was found to be replicable using a passive flexible structure. Flaps of constant cross section and varying lengths were put on the robotic vehicle to conduct a systematic parametric study. Robojelly's swimming performance was tested with and without a flap. This revealed a thrust increase 1340% with the addition of a flap.  Velocity field measurements were performed using planar Time Resolved Digital Particle Image Velocimetry (TRDPIV) to analyze the change in vortex structures as a function of flap length.  The robot input parameters stayed constant over the different configurations tested thus maintaining a near constant power consumption. Non-dimensional circulation results show a dependence on flap kinematics and geometry. The robot was approximated as a series of pitching panels circularly oriented around its apex. The first circulation peak of the pitching panel approximation revealed a normalized standard deviation of 0.23. A piston apparatus was designed and built to test different flexible margin configurations. This apparatus allow the isolation of the flap parameters and remove the uncertainties coming from the robotic vehicle. / Ph. D.

Diabetes patients are increasingly suffering from wound-healing impairments coupled with other severe medical conditions. High levels of oxidative stress and proinflammatory cytokines are predominant pathologic characteristics of diabetic wounds. Applying microRNA (miRNA) in gene therapy often successfully treats various disorders by generating highly specific bio-responses. Studies revealed that microRNA146a (miR146a), a regulator of inflammation, is downregulated in diabetic wounds. However, miR203 is found to be highly-expressed in diabetic foot ulcers in positive correlation with disease severity. We have synthesized Cerium oxide nanoparticles (CNPs) since they are used in various biomedical applications due to their high enzyme-mimetic activities and reactive oxygen species (ROS) scavenging abilities. In the present work, two studies are presented. The first study examines whether CNPs can stabilize and protect surface-bound miR146a from oxidative damage under excess ROS exposure. We assessed the relative performance of a miR146a-conjugated CNP formulation (CNP-miR146a) by comparing against two other nanoparticle compositions widely used in nanomedicine: gold and silica. Results from material characterizations and electroanalytical studies demonstrated stable catalytic responses of CNP-miR146a toward ROS and CNPs' ability to protect miR146a from oxidative damage, while the other formulations were relatively ineffective. In the second study, different copper concentrations were incorporated into the nanocrystalline CNP lattice. The defects created in CNPs and the differences between produced formulations were evaluated through XPS, TEM, and XRD. An electrochemical biosensor was developed using the electrochemical charge transfer properties of copper-modified CNP formulation to determine miR203 concentration. Biosensor function was assessed by electroanalytical studies: showing a detection range from 10 μM down to the detection limit 1.73 fM. The present studies demonstrated that the surface catalytic activities of CNP formulations can mediate preservation of miRNA functional integrity and allow quantitative detection of miRNA in an electrochemical biosensor platform. Together, these studies suggest unique value for CNPs in miRNA-based biotechnologies.

Implantable medical devices face numerous complications when interfacing with soft tissue, and are plagued by negative responses from host tissue. One such class devices are percutaneous osseointegrated prosthetics (POP). POP consist of a bone anchored titanium post that extrudes through the skin and attaches to an external prosthetic. Compared to the traditional socket interface, POPs offer better stability, limb functionality, and osseoperception for both upper and lower prosthetic limbs. Although the POP surgery technique is well established, the main disadvantage to this technology remains the titanium (Ti) - skin interface. Some of the complications that can arise include epithelial downgrowth, mechanical tearing, and infection. Various types of coatings, surface structure, and antibiotic release technologies have been used to coat Ti in an effort to mitigate POP's associated obstacles, but these methods have failed to translate into published clinical studies and mainstream medical use. One potential solution may be to mimic an interface already found in the human body, the fingernail-skin interface, which is infection-free and mechanically stable. The same keratins that make up the cortex of human hair fibers are found in the fingernail. These cortical human hair keratins can be extracted and purified, and fingernail-specific dimeric complexes coated onto Ti surfaces using silane coupling chemistry. Keratin has been used in other studies for its cell adhesion and differentiation properties, and it has been suggested that the Leu-Asp-Val (LDV) amino acid motif is the primary site responsible for cellular attachment. In the present work, keratins extracted from human hair fibers and recombinant keratin nanomaterials (KN) were used to create biomimetic coatings on silanized Ti surfaces. These coatings were characterized and investigated for surface topography, elemental composition, cell adhesion motifs, and cell adhesion. Both keratin substrates showed the ability to create uniform coatings that retain a protein conformation that exhibits cell adhesion motifs. The coatings exhibit the ability to support cell adhesion of both epithelial and connective tissue cells. Application of fluid shear stress was used to test the mechanical adhesion strength of cells on keratin coatings. The structure, biochemical stability and sustained cellular adhesion of these coatings support keratin's capacity to provide a stable interface between POPs and skin. Side-by-side studies of extracted and recombinant keratins reveals that the recombinant form of these materials may provide distinct advantages for their use in POP devices. Overall, this study confirmed that a uniform, silane-coupled keratin coating was feasible. We demonstrated the substrates contain a biological function in terms of cellular adhesion and phenotypic changes in skin-relevant cells. These results support the biomimetic function of keratin on silanized Ti, which may provide a suitable coating to translate percutaneous medical device coating applications toward clinical use. / Ph. D. / Implantable medical devices face numerous complications when interfacing with soft tissue, and are plagued by negative responses from host tissue. One such class devices is percutaneous osseointegrated prosthetics (POP). POP consist of a bone anchored titanium post that extrudes through the skin and attaches to an external prosthetic. Compared to the traditional socket interface, POPs offer better stability, limb functionality, and osseoperception for both upper and lower prosthetic limbs. Although the POP surgery technique is well established, the main disadvantage to this technology remains the titanium (Ti) - skin interface. Some of the complications that can arise include epithelial downgrowth, mechanical tearing, and infection. Various types of coatings, surface structure, and antibiotic release technologies have been used to coat Ti in an effort to mitigate POP’s associated obstacles, but these methods have failed to translate into published clinical studies and mainstream medical use. One potential solution may be to mimic an interface already found in the human body, the fingernail-skin interface, which is infection-free and mechanically stable. The same keratins that make up the cortex of human hair fibers are found in the fingernail. These cortical human hair keratins can be extracted and purified, and fingernail-specific dimeric complexes coated onto Ti surfaces using silane coupling chemistry. Keratin has been used in other studies for its cell adhesion and differentiation properties, and it has been suggested that the Leu-Asp-Val (LDV) amino acid motif is the primary site responsible for cellular attachment. In the present work, keratins extracted from human hair fibers and recombinant keratin nanomaterials (KN) were used to create biomimetic coatings on silanized Ti surfaces. These coatings were characterized and investigated for surface topography, elemental composition, cell adhesion motifs, and cell adhesion. Both keratin substrates showed the ability to create uniform coatings that retain a protein conformation that exhibits cell adhesion motifs. The coatings exhibit the ability to support cell adhesion of both epithelial and connective tissue cells. Application of fluid shear stress was used to test the mechanical adhesion strength of cells on keratin coatings. The structure, biochemical stability and sustained cellular adhesion of these coatings support keratin’s capacity to provide a stable interface between POPs and skin. Side-by-side studies of extracted and recombinant keratins reveals that the recombinant form of these materials may provide distinct advantages for their use in POP devices. Overall, this study confirmed that a uniform, silane-coupled keratin coating was feasible. We demonstrated the substrates contain a biological function in terms of cellular adhesion and phenotypic changes in skin-relevant cells. These results support the biomimetic function of keratin on silanized Ti, which may provide a suitable coating to translate percutaneous medical device coating applications toward clinical use.

As the technology of robotics intelligence advances, and new application areas for mobile robots increase, the need for alternative fundamental locomotion mechanisms for robots that allow them to maneuver into complex unstructured terrain becomes critical. In this research we present a novel locomotion mechanism for mobile robots inspired by the motility mechanism of certain single celled organisms such as amoebae. Whole Skin Locomotion (WSL), as we call it, works by way of an elongated toroid which turns itself inside out in a single continuous motion, effectively generating the overall motion of the cytoplasmic streaming ectoplasmic tube in amoebae. This research presents the preliminary analytical study towards the design and development of the novel WSL mechanism. In this thesis we first investigate how amoebas move, then discuss how this motion can be replicated. By applying the biological theories of amoeboid motility mechanisms, different actuation models for WSL are developed including the Fluid Filled Toroid (FFT) and Concentric Solid Tube (CST) models. Then, a quasi-static force analysis is performed for the CST model and parametric studies for design, including power efficiency and force transition characteristics, are presented. / Master of Science

Current orthopedic implant technologies used suffer from slow rates of osseointegration, short lifetime, and lack of mechanical integrity as a result of poorly controlled cell-surface interactions. Recent biologically-inspired surface strategies (biomimetic) have focused on mimicking the biofunctionality of the extracellular matrix (ECM) by using short, adhesive oligopeptides, such as arginine-glycine-aspartic acid (RGD) present in numerous ECM components. However, these strategies have yielded mixed results in vivo and marginal bone healing responses. The central goal of this dissertation project was to engineer bioactive surfaces that specifically target integrin receptors important for osteogenic functions in order to improve bone tissue repair. In order to create integrin-specific interfaces, integrin-specific ligands reconstituting the fibronectin (FN) secondary/tertiary structure were first engineered and functionalized on material surfaces using several robust presentation schemes. We demonstrated that FN-mimetic-functionalized surfaces that directed α5β1 binding enhanced osteoblast and stromal cell integrin binding and adhesion, osteogenic signaling, and osteoblastic differentiation compared to various other RGD-based ligand-functionalized surfaces. Next, we investigated the effect of integrin-specific biointerfaces to modulate bone healing in a rat tibia implant bone model. We demonstrated, using a robust polymer brush system, that bioactive coatings on titanium implants that conferred high α5β1 integrin specificity in vitro enhanced bone formation and implant integration in vivo. Moreover, we showed that integrin specificity can be engineered using different immobilization schemes, including clinically-relevant ligand dip-coating, and promote the same robust in vivo effect. Furthermore, we investigate the synergistic roles of integrin specificity and ligand clustering on cell response by engineering biointerfaces presenting trimeric and pentameric "heads" of FNIII7-10 with nanoscale spacing. Integrin-specific ligand clustering supported α5β1-specific binding and cell adhesion and enhanced implant osseointegration in vivo compared to monovalent FNIII7-10 or non-functionalized biointerfaces. In summary, the FN-mimetic integrin-specific biointerfaces engineered in this thesis provide a clinically-relevant material surface strategy to modulate tissue healing responses. In addition, these results contribute to our greater understanding of how two specific material design parameters, integrin binding specificity and clustered ligand presentation, contribute individually and synergistically toward directing cell and tissue function.

Insects posses an incredible ability to navigate their environment at high speed, despite having small brains and limited visual acuity. Through selective pressure they have evolved computationally efficient means for simultaneously performing navigation tasks and instantaneous control responses. The insect’s main source of information is visual, and through a hierarchy of processes this information is used for perception; at the lowest level are local neurons for detecting image motion and edges, at the higher level are interneurons to spatially integrate the output of previous stages. These higher level processes could be considered as models of the insect's environment, reducing the amount of information to only that which evolution has determined relevant. The scope of this thesis is experimenting with biologically inspired visual control of flying robots through information processing, models of the environment, and flight behaviour. In order to test these ideas I developed a custom quadrotor robot and experimental platform; the 'wasp' system. All algorithms ran on the robot, in real-time or better, and hypotheses were always verified with flight experiments. I developed a new optical flow algorithm that is computationally efficient, and able to be applied in a regular pattern to the image. This technique is used later in my work when considering patterns in the image motion field. Using optical flow in the log-polar coordinate system I developed attitude estimation and time-to-contact algorithms. I find that the log-polar domain is useful for analysing global image motion; and in many ways equivalent to the retinotopic arrange- ment of neurons in the optic lobe of insects, used for the same task. I investigated the role of depth in insect flight using two experiments. In the first experiment, to study how concurrent visual control processes might be combined, I developed a control system using the combined output of two algorithms. The first algorithm was a wide-field optical flow balance strategy and the second an obstacle avoidance strategy which used inertial information to estimate the depth to objects in the environment - objects whose depth was significantly different to their surround- ings. In the second experiment I created an altitude control system which used a model of the environment in the Hough space, and a biologically inspired sampling strategy, to efficiently detect the ground. Both control systems were used to control the flight of a quadrotor in an indoor environment. The methods that insects use to perceive edges and control their flight in response had not been applied to artificial systems before. I developed a quadrotor control system that used the distribution of edges in the environment to regulate the robot height and avoid obstacles. I also developed a model that predicted the distribution of edges in a static scene, and using this prediction was able to estimate the quadrotor altitude.

There is an ongoing interest in analyzing the flow characteristics of swimming fish. Biology has resulted in some very efficient motions and formulating these motions is of interest to engineers. One such theory was written by Dr. William Vorus and Dr. Brandon Taravella involving ideal efficiency. It is therefore interesting to test the calculations to see if it is possible to design a motion that can create thrust without necessarily creating vorticity. The computational fluid dynamics software of ANSYS Fluent was used to calculate the resulting flow field of the eel motion to compare with the theoretical values.