Gih-Keong Lau

Dragonflies' exceptional climb capability plays a crucial role in essential survival behaviors such as predation and evasion. However, the climb performance in aerial vehicles remains relatively weak, posing challenges in achieving efficient and stable vertical ascent. To investigate the body motion patterns and wing flapping strategies during climbing flight, this study first conducts biological observations to obtain basic biological characteristics and key climbing parameters. Based on the observational data, numerical simulations are performed to explore aerodynamic performance during ascent. The flow field structures are analyzed through flow visualization. The results show that the dragonfly's climbing process can be divided into four stages, with the average flapping frequency being 28.0 Hz and the total duration being approximately 0.82 s. The ascent is predominantly vertical, with slight backward flight and turning near the end. The angle of attack is essential for regulating flight posture and speed. Meanwhile, by adjusting the flapping plane tilt angle, deviation angle, and amplitude, the dragonfly maintains high lift and stable climbing through the synergistic effects of typical vortex structures such as leading-edge vortices. During the rapid-ascent phase, a significant pressure difference occurs on the wing surfaces along with delayed stall, and rotational circulation is also formed, both of which are key mechanisms for enhanced aerodynamic force generation. Furthermore, the dragonfly dynamically balances lift and flight maneuverability by modulating the phase difference between the forewings and hindwings, with their interaction resulting in a maximum aerodynamic force increase of 25.1%, highlighting the sophisticated aerodynamic strategies during climbing flight.

This study presents a flapping-wing micro aerial vehicle designed for three-dimensional trajectory tracking. The design features center-of-gravity-based steering, three actuators, and omits a dedicated yaw actuator, time-varying wingbeat patterns, and high-frequency actuator switching. To investigate the dynamics of the proposed design, a modified nonlinear controllability framework was developed for theoretical analysis, and a series of flight experiments were conducted to validate the analytical results. The analysis indicates that, despite the absence of a dedicated yaw actuator, the yaw torque required for three-dimensional maneuvering arises from center-of-gravity shifts, asymmetric wing aerodynamics, and an intrinsic feedback mechanism that stabilizes the angle difference between vehicle's yaw angle and trajectory rotation angle. Furthermore, the yaw dynamics exhibit nonminimum phase behaviors in response to center-of-gravity movement. Based on these findings, a multi-tier control architecture is proposed to accommodate these unique dynamics for autonomous flight. Experimental results demonstrate that the proposed design can perform three-dimensional trajectory tracking, with attitude errors below 5° and altitude deviations within 10 cm in most cases.

Ornithopters are bird‐like flapping‐wing robots. Only small ornithopters can hover, with long endurance at hummingbird size. Could larger ornithopters be improved further to hover longer? This paper reviews and examines the drive and power of hovering ornithopters, and elastic means of energy or thrust boosters. While the rotation of flexible wings enhance the thrust generation, two‐winged ornithopters did not scale up well because of higher disk loading. In comparison, the X‐winged or multiple‐V‐winged ornithopters enjoy a lower disk loading by beating multiple wings slower, at a smaller stroke angle or a longer span. Further, the clap‐and‐fling interaction of V and X‐wings boosts the thrust generation. Future works can explore the wing flexibility and morphology change to improve the hoverability and flight agility of ornithopters.

Fast drumming presents a speed challenge to many robotic arms. To simultaneously meet the needs for speed, stroke, and force, we proposed and tested a new double-saddle dielectric elastomer actuator (DEA) as the artificial biceps for driving a lightweight robotic drummer. This work finds that fast force induction is instrumental to the fast drumming by a DEA-driven drumstick. While a pure-shear series of DEA under a large pre-stretch and lateral reinforcement can generate a large isotonic stroke, it is not fast in isometric force induction. Instead, this work found a double-saddle DEA, which is a degenerated two-segment pure-shear DEA with a middle laterally buckled beam, managed to induce a faster isometric force at a 0.5 s down time constant when its ultimate actuation tapered due to substantial strain stiffening effect. As such, this double-saddle DEA-driven drummer suffered less from the dynamic stroke decrement than the 8-segment DEA did. A 4.4 g DEA managed to swing freely a 7.8 g drumstick up to nearly 60° at 6 kV and a maximum tip speed of up to 0.45 m/s. The drumming frequency upon 5.5 kV activation was up to 2–2.5 Hz. In comparison, an 8-segment pure-shear DEA-driven drummer failed to reach the drum when pulsed at 2 Hz. Interestingly, a slower stroke excited multiple drumbeats due to secondary bounces. To achieve a faster drumming to match a human drummer, it is foreseen that a harder dielectric elastomer material will help drive a faster soft robot.

A ventilated acoustic metasurface consisting of a membrane covered with a combination of different depth sub-chambers is proposed. It can achieve at least a 5 dB sound insulation acoustic performance in the wide frequency range from 100 to 1700 Hz, in particular a 10 dB noise reduction in the range from 100 to 200 Hz and from 437.4 to 1700 Hz, which can therefore cover the low-frequency range of the environmental noise. The physical mechanism of membrane-acoustic coupling for noise reduction in the low-frequency range is further explored.

Road noise varies with the vehicles’ load and speed, ranging from low to medium frequencies. The existing absorbers, including acoustic metamaterials, cannot fully mitigate the variable noise due to limited bandwidth. To target the variable noise but not occupy much space, this paper presents a tunable acoustic absorber that can transform from Helmholtz resonance to membrane resonance. This tunable absorber consist of double micro‐slit layers, namely an electro‐adhesive membrane and a flexible plate spaced closely in front of a shallow cavity. The boundary‐layer thick gap between the two forms a network of lateral orifices to enhance the acoustic damping at medium frequencies. When the gap closes under electrostatic attraction force, the membrane and plate zipped together and resonated like a drum at a much lower frequency. This transformation can shift the resonant frequency down by 1.6 octaves from 898 hetz (Hz) to 281 Hz, thanks to the structural resonance of the zipped parts. Interestingly, the peak absorption coefficient at the lowest tuned 281 Hz was as high as 0.93. Such distributed tunability based on a fault‐tolerant electro‐adhesive membrane will also be applicable to other acoustic meta‐materials that embody the components of micro‐perforated or micro‐slit panels.

Most modern high-rise buildings' facades use glasses for esthetics, daylight, and better environmental view through them. However, with the increasing use of a larger area of transparent glasses as walls, the visual privacy preservation of the indoors and heat energy loss through the windows are becoming a rising concern. Recent studies showed that nearly half of the energy consumed in a building goes to heating and air conditioning while approximately 40% of this energy is lost through windows. Windows with tunable optical properties that are generically termed “tunable windows or smart windows or switchable glass” are perceived as a potential solution for these problems. An optically tunable window can adjust the amount of daylighting passing through it, control the heat radiation, and/or change the transparency of the glasses for visual privacy preservation of indoors. Electrochromic glasses, polymer dispersed liquid crystal glasses, and suspended particle devices are available as commercial tunable windows but their high cost, limited optical performance, reliability, and operational complexity are hindering the widespread adaptation. Therefore, several other technologies for low-cost actively tunable windows capable of actively adjusting transparency are increasingly explored. However, most of such new technologies, working based on various optical principles, do not fulfill all the requirements of tunable windows. For instance, some can tune optical transmittance but do not affect energy transmission, and some can adjust heat radiation transmission but has a limited change in visual appearances. To fully take the advantage of the strengths as well as recognize the limitations of such emerging technologies, their optical principles need to be understood in-depth. Here, we review the recent developments in transmittance tunable windows by categorizing them based on the optics involved, namely, light absorption, reflection, and scattering. This in-depth review comprehensively discusses how the tunable window technologies compare to each other and offers insight into how their performance can be improved in the future.

Harder dielectric elastomers are structurally stronger than soft ones but hardly actuate under electrostatic pressure. It remains unclear which property limits harder dielectric elastomers from producing as large actuation as soft ones. This study shows that the change in stretch limit rather than modulus bounds the ultimate dielectric actuation. A larger ultimate stretch warrants a pre-stretched membrane to further thin down and expand biaxially relative to the pre-stretch state. Theory and experiments showed that pre-stretch alters the maximum stretchability and ultimate dielectric actuation. An optimal pre-stretch helps maximize the ultimate actuation close to 2 times radially so long as the corresponding total stretch did not exceed the elastomer’s fracture limit. Hence, a stretchy harder dielectric elastomer stands a chance of producing as large ultimate actuation as soft ones do. To validate the theory, we tested three commercial grades of styrene–ethylene/butylene–styrene copolymers (SEBS) organogels with different oil content and shore hardness. The tests showed that a 20 A-hardness organogel of low toughness actuated less ultimately, but a 10 A-hardness organogel of greater toughness actuated as much as an 8 A-hardness organogel does. This 10 A-hardness SEBS organogel showed an ultimate elongation of 1192%, and 3.1 times the shear modulus and seven times the tensile strength of 3 M VHB 4910. When being optimally pre-stretched for 3.5 times radially, it produced a substantial hyperelastic dielectric actuation up to 104% active areal strain at 277 MV m ⁻¹ , comparable to the actuation generated by the over-stiffened VHB DEAs. Unlike VHB DEAs, these harder SEBS DEAs generated hyperelastic actuation free from severe viscoelastic drift. The high performance of SEBS organogels can potentially meet the need for forceful artificial muscles to drive bio-inspired robots.

Dielectric elastomer minimum energy structures (DEMESs) are useful as low-force robotic grippers; they can sweep a large angle but carry not much load. It was a design dilemma to reinforce the benders without compromising the stroke angle. As a stronger variant of DEMES, a dielectric elastomer (DE) finger can unbend the ‘phalanges’ of a load beam upon activation of the ‘intrinsic muscles’ of the dielectric elastomer actuator (DEA). The DE finger used a uniform tendon hood that raises the tension center of a single-layered DEA and thus enhances the moment generation and load capacity. In this work, we further optimize the structural design of a slender DE finger by mimicking the human thumb profile. This thumb-inspired DE finger has a tapered load beam for hood shaping of multi-layered DEAs with a blunter fingertip. This thumb-up profile greatly enhances the passive lift strength (against a tip weight) by 54% as compared to the earlier rectangular design, at the cost of a 13% reduction in the active stroke. Further, it exploited the axial stiffness to achieve an order-greater pull strength as compared to the lift strength. Finally, the optimized DEMES grippers carried a payload well exceeding the lift strength; they managed to pick an apple of nearly ten times the gripper weight. In addition, a foot of three DE toes hung upside-down to a branch of a horizontal tube while supporting a payload of close to ten times the foot’s weight.