Research ArticleBIOMIMETICS

A biologically inspired, flapping-wing, hybrid aerial-aquatic microrobot

See allHide authors and affiliations

Science Robotics  25 Oct 2017:
Vol. 2, Issue 11, eaao5619
DOI: 10.1126/scirobotics.aao5619
  • Fig. 1 Robot design, component fabrication, and assembly.

    (A) An existing 85-mg robot was used to investigate underwater stability. (B) The improved 175-mg robot consisted of two symmetric halves, a central gas collection chamber with a sparker plate, four balance beams, and buoyant outriggers. (C) Exploded view of robot assembly. Scale bar, 1 cm (B and C). (D) Mating feature of the titanium balance T-beam. Scale bar, 500 μm. (E) Exploded view of gas collection chamber assembly. Scale bar, 5 mm. (F) Microscopic image illustrating an array of porous openings on the chamber’s titanium top plate. Scale bar, 500 μm. (G) The sparking plate consists of a pair of stainless steel plates and a copper sparker. Scale bar, 4 mm. (H) Microscopic image of the sparker electrodes. Scale bar, 100 μm.

  • Fig. 2 Demonstration of aerial-aquatic locomotion and transition.

    (A) The robot is capable of aerial hovering, air-to-water transition, swimming, water-to-air transition, impulsive takeoff, and landing. (B) Composite image of a hovering robot. (C) Composite image of the robot transitioning from air to water. (D) Composite image of the robot swimming to the water surface. (E) Images of the robot gradually emerge from the water surface by capturing gas from electrolysis. (F) Composite image of robot takeoff and landing. Scale bars, 1 cm.

  • Fig. 3 Simulations and experiments of robot swimming stability.

    (A) Simulation of the robot center of mass motion when it is driven at 11 Hz. The color scale represents vehicle speed and has units of millimeter per second. The flapping frequency is 11 Hz. (B) Robot center of mass velocity. (C) Robot body rotation. φ, θ, and ψ represent the body yaw, pitch, and roll motion, respectively. (A to C) Results of the same simulation. (D) Composite image of an unstable swimming robot operating at 5 Hz. (E) This robot experiences notable body pitching (14.8°) when it flaps wings at 5 Hz. (F) Composite image of an upright stable robot ascending to the water surface. (G) The robot pitching amplitude reduces to 9.4° when swimming frequency increases to 11 Hz. Scale bars, 1 cm (D to G). (H) Simulation results of wing stroke (α) and pitch (β) amplitude and relative phase (δ) as functions of flapping frequency. (I) Experimental and simulation comparison of robot pitch amplitude as a function of flapping frequency. (J) Experimental and simulation comparison of robot ascent speed as a function of flapping frequency. (H to J) Red and blue colors distinguish regions that are either stable or unstable, respectively. Both experiments and simulations show that the robot is unstable when the flapping frequency is lower than 9 Hz.

  • Fig. 4 Robot surface tension experiments.

    (A) Illustration of the experimental setup. The robot is mounted on a capacitive force sensor and is slowly lowered into water or pulled out of water. (B) Force trace when the robot is lowered into soapy water. (C) Force trace as the robot is lowered into tap water. (D) Picture of the robot and force trace immediately before one of its balance beams (red circle) pops out of the water surface. (E) Picture of the robot and force trace immediately after one of its balance beams (red circle) pops out of the water surface. (D) and (E) are taken 0.33 s apart. (F) Force trace as the robot is pulled out of soapy water. (G) Force trace as the robot is pulled out of tap water. Scale bars, 1 cm (A, D, and E).

  • Fig. 5 Impulsive takeoff from the water surface.

    (A) Sparker plate input voltage and current when a visible spark is generated. (B) Pressure profile within the chamber upon oxyhydrogen ignition. A reinforced chamber without micro-openings experiences higher pressure than one with micro-openings. (C) Image sequence comparison of initial robot takeoff. For the robot without chamber micro-openings (left), the detonation cracks the chamber top plate and detaches a robot balance beam and wing. For the chamber with micro-openings (right), gas and water are released upon ignition, and the robot remains undamaged. (D) Overlaid image comparison of robot takeoff. A robot without chamber micro-openings experiences substantial body rotation and has a higher takeoff speed (left). A robot with chamber micro-openings maintains upright stability and has lower takeoff speed (right). Scale bars, 1 cm.

Supplementary Materials

  • robotics.sciencemag.org/cgi/content/full/2/11/eaao5619/DC1

    Text S1. Electrolytic plate geometry and efficiency.

    Text S2. Effect of micro-openings on gas capture.

    Text S3. Robot stability near the water surface.

    Text S4. Effect of micro-openings on takeoff.

    Text S5. Derivation of dynamical model.

    Text S6. Robot tracking.

    Text S7. Simplified model of robot passive upright stability.

    Fig. S1. Material selection of sparker and electrolytic plates, and plate geometry influence on water resistance during electrolysis.

    Fig. S2. Swimming demonstration of the new robot design.

    Fig. S3. Robot water entry from different orientations.

    Fig. S4. Surface tension influence on height of the gas collection chamber.

    Fig. S5. Robot stability near the water surface.

    Fig. S6. Influence of micro-openings on takeoff speed.

    Fig. S7. Influence of micro-openings on ignition pressure and takeoff speed.

    Fig. S8. Comparison of flapping kinematics before and after impulsive takeoff.

    Fig. S9. Robot liftoff demonstration before and after impulsive takeoff.

    Fig. S10. Experimental setup.

    Fig. S11. Dynamical model and motion-tracking method.

    Fig. S12. Robot stability during freefall and swimming.

    Table S1. Properties of robot components.

    Table S2. Model parameter values.

    Movie S1. Demonstration of robot aerial hover.

    Movie S2. Demonstration of robot air-water transition.

    Movie S3. Demonstration of robot swimming and emergence of robot wing from the water surface.

    Movie S4. Demonstration of robot impulsive takeoff and landing.

    Movie S5. Comparison of robot underwater stability with different flapping frequencies.

    Movie S6. Comparison between robot swimming experiment and simulation.

    Movie S7. Measurement of surface tension force on a robot during water-to-air transition.

    Movie S8. Comparison of robot takeoff with or without micro-openings on gas collection chamber.

    Movie S9. Detonation pressure measurement and robot takeoff.

    References (3235)

  • Supplementary Materials

    Supplementary Material for:

    A biologically inspired, flapping-wing, hybrid aerial-aquatic microrobot

    Yufeng Chen,* Hongqiang Wang, E. Farrell Helbling, Noah T. Jafferis, Raphael T. Zufferey, Aaron Ong, Kevin Ma, Nicholas Gravish, Pakpong Chirarattananon, Mirko Kovac, Robert J. Wood*

    *Corresponding author. Email: rjwood{at}eecs.harvard.edu (R.J.W.); yufengchen{at}seas.harvard.edu (Y.C.)

    Published 25 October 2017, Sci. Robot. 2, eaao5619 (2017)
    DOI: 10.1126/scirobotics.aao5619

    This PDF file includes:

    • Text S1. Electrolytic plate geometry and efficiency.
    • Text S2. Effect of micro-openings on gas capture.
    • Text S3. Robot stability near the water surface.
    • Text S4. Effect of micro-openings on takeoff.
    • Text S5. Derivation of dynamical model.
    • Text S6. Robot tracking.
    • Text S7. Simplified model of robot passive upright stability.
    • Fig. S1. Material selection of sparker and electrolytic plates, and plate geometry influence on water resistance during electrolysis.
    • Fig. S2. Swimming demonstration of the new robot design.
    • Fig. S3. Robot water entry from different orientations.
    • Fig. S4. Surface tension influence on height of the gas collection chamber.
    • Fig. S5. Robot stability near the water surface.
    • Fig. S6. Influence of micro-openings on takeoff speed.
    • Fig. S7. Influence of micro-openings on ignition pressure and takeoff speed.
    • Fig. S8. Comparison of flapping kinematics before and after impulsive takeoff.
    • Fig. S9. Robot liftoff demonstration before and after impulsive takeoff.
    • Fig. S10. Experimental setup.
    • Fig. S11. Dynamical model and motion-tracking method.
    • Fig. S12. Robot stability during freefall and swimming.
    • Table S1. Properties of robot components.
    • Table S2. Model parameter values.
    • Legends for movies S1 to S9
    • References (3235)

    Download PDF

    Other Supplementary Material for this manuscript includes the following:

    • Movie S1 (.mp4 format). Demonstration of robot aerial hover.
    • Movie S2 (.mp4 format). Demonstration of robot air-water transition.
    • Movie S3 (.mp4 format). Demonstration of robot swimming and emergence of robot wing from the water surface.
    • Movie S4 (.mp4 format). Demonstration of robot impulsive takeoff and landing.
    • Movie S5 (.mp4 format). Comparison of robot underwater stability with different flapping frequencies.
    • Movie S6 (.mp4 format). Comparison between robot swimming experiment and simulation.
    • Movie S7 (.mp4 format). Measurement of surface tension force on a robot during water-to-air transition.
    • Movie S8 (.mp4 format). Comparison of robot takeoff with or without microopenings on gas collection chamber.
    • Movie S9 (.mp4 format). Detonation pressure measurement and robot takeoff.

    Files in this Data Supplement:

Navigate This Article