Large-scale automated investigation of free-falling paper shapes via iterative physical experimentation

Free-falling paper shapes exhibit rich, complex and varied behaviours that are extremely challenging to model analytically. Physical experimentation aids in system understanding, but is time-consuming, sensitive to initial conditions and reliant on subjective visual behavioural classification. In this study, robotics, computer vision and machine learning are used to autonomously fabricate, drop, analyse and classify the behaviours of hundreds of shapes. The system is validated by reproducing results for falling discs, which exhibit four falling styles: tumbling, chaotic, steady and periodic. A previously determined mapping from a non-dimensional parameter space to behaviour groups is shown to be consistent with these new experiments for tumbling and chaotic behaviours. However, steady or periodic behaviours are observed in previously unseen areas of the parameter space. More complex hexagon, square and cross shapes are investigated, showing that the non-dimensional parameter space generalizes to these shapes. The system highlights the potential of robotics for the investigation of complex physical systems, of which falling paper is one example, and provides a template for future investigation of such systems. The dynamics of paper shapes in free fall are still not fully understood, despite being discussed for more than 150 years. Collecting large amounts of data has the potential to give us new insights and a robotics system could generate and analyse data in large quantities.

[1]  Kapil Varshney,et al.  Unsteady aerodynamic forces and torques on falling parallelograms in coupled tumbling-helical motions. , 2013, Physical review. E, Statistical, nonlinear, and soft matter physics.

[2]  Helmut Hauser,et al.  Towards a theoretical foundation for morphological computation with compliant bodies , 2011, Biological Cybernetics.

[3]  J. A. Fox,et al.  Experimental Investigation of the Stability of Hagen‐Poiseuille Flow , 1968 .

[4]  Tao Li,et al.  Information processing via physical soft body , 2015, Scientific Reports.

[5]  S. West,et al.  The Illusion of Invariant Quantities in Life Histories , 2005, Science.

[6]  Bruce G Buchanan,et al.  Automating Science , 2009, Science.

[7]  Mark Peplow,et al.  Organic synthesis: The robo-chemist , 2014, Nature.

[8]  Stefan Grimm,et al.  High-throughput isolation of ultra-pure plasmid DNA by a robotic system , 2006, BMC biotechnology.

[9]  Christopher H. Bryant,et al.  Functional genomic hypothesis generation and experimentation by a robot scientist , 2004, Nature.

[10]  Fumiya Iida,et al.  The trade-off between morphology and control in the co-optimized design of robots , 2017, PloS one.

[11]  J. A. Hartigan,et al.  A k-means clustering algorithm , 1979 .

[12]  Tim Chapman.  Lab automation and robotics: Automation on the move , 2003, Nature.

[13]  J. Magnaudet,et al.  On the zigzag dynamics of freely moving axisymmetric bodies , 2005 .

[14]  Fumiya Iida,et al.  Model-Free Design Optimization of a Hopping Robot and Its Comparison With a Human Designer , 2018, IEEE Robotics and Automation Letters.

[15]  Fumiya Iida,et al.  Evolutionary Developmental Robotics: Improving Morphology and Control of Physical Robots , 2017, Artificial Life.

[16]  R. Dimitrakopoulos,et al.  High-order Stochastic Simulation of Complex Spatially Distributed Natural Phenomena , 2010 .

[17]  Elisha Moses,et al.  From Flutter to Tumble: Inertial Drag and Froude Similarity in Falling Paper , 1998 .

[18]  Z. J. Wang,et al.  The kinematics of falling maple seeds and the initial transition to a helical motion , 2011 .

[19]  Hod Lipson,et al.  Distilling Free-Form Natural Laws from Experimental Data , 2009, Science.

[20]  Daryl B. Simons,et al.  The behavior of large particles falling in quiescent liquids , 1969 .

[21]  Shiyi Chen,et al.  Experimental study of freely falling thin disks: Transition from planar zigzag to spiral , 2011 .

[22]  Changqiu Jin,et al.  Numerical study of the unsteady aerodynamics of freely falling plates , 2008 .

[23]  Robert M. May,et al.  Simple mathematical models with very complicated dynamics , 1976, Nature.

[24]  Maria Liakata,et al.  Towards Robot Scientists for autonomous scientific discovery , 2010, Automated experimentation.

[25]  Eva Kanso,et al.  Coins falling in water , 2013, Journal of Fluid Mechanics.

[26]  E Mjolsness,et al.  Machine learning for science: state of the art and future prospects. , 2001, Science.

[27]  K. Mikaelian.  Analytic Approach to Nonlinear Rayleigh-Taylor and Richtmyer-Meshkov Instabilities , 1998 .

[28]  F. Iida,et al.  Physics driven behavioural clustering of free-falling paper shapes , 2019, PloS one.

[29]  David E. Goldberg,et al.  Genetic algorithms and Machine Learning , 1988, Machine Learning.

[30]  Felice C. Frankel,et al.  Big data: Distilling meaning from data , 2008, Nature.

[31]  Jorge Nocedal,et al.  Optimization Methods for Large-Scale Machine Learning , 2016, SIAM Rev..

[32]  Tom Schaul,et al.  Unifying Count-Based Exploration and Intrinsic Motivation , 2016, NIPS.

[33]  John Rieffel,et al.  Growing and Evolving Soft Robots , 2014, Artificial Life.

[34]  R. Norberg.  AUTOROTATION, SELF‐STABILITY, AND STRUCTURE OF SINGLE‐WINGED FRUITS AND SEEDS (SAMARAS) WITH COMPARATIVE REMARKS ON ANIMAL FLIGHT , 1973 .

[35]  Franco Nori,et al.  Chaotic dynamics of falling disks , 1997, Nature.

[36]  Beric W. Skews,et al.  Autorotation of rectangular plates , 1990 .

[37]  James Clerk Maxwell,et al.  The Scientific Papers of James Clerk Maxwell: On a particular case of the descent of a heavy body in a resisting medium , 2011 .

[38]  Maria Liakata,et al.  An Integrated Laboratory Robotic System for Autonomous Discovery of Gene Function , 2010 .

[39]  Amanda Clare,et al.  An ontology for a Robot Scientist , 2006, ISMB.

[40]  F. Iida,et al.  Morphological Evolution of Physical Robots through Model-Free Phenotype Development , 2015, PloS one.

[41]  Kenneth Showalter,et al.  Nonlinear Chemical Dynamics: Oscillations, Patterns, and Chaos , 1996 .

[42]  Filip De Turck,et al.  #Exploration: A Study of Count-Based Exploration for Deep Reinforcement Learning , 2016, NIPS.

[43]  Shiyi Chen,et al.  Experimental investigation of freely falling thin disks. Part 2. Transition of three-dimensional motion from zigzag to spiral , 2013, Journal of Fluid Mechanics.

[44]  Lionel Vincent,et al.  Holes stabilize freely falling coins , 2016, Journal of Fluid Mechanics.

[45]  J. de Boer,et al.  Robot-scientists will lead tomorrow's biomaterials discovery , 2018, Current Opinion in Biomedical Engineering.

[46]  S. Xu,et al.  Influence of aspect ratio on tumbling plates , 2013, Journal of Fluid Mechanics.

[47]  Z. Jane Wang,et al.  Analysis of transitions between fluttering, tumbling and steady descent of falling cards , 2005, Journal of Fluid Mechanics.

[48]  D. Fan,et al.  A robotic Intelligent Towing Tank for learning complex fluid-structure dynamics , 2019, Science Robotics.

[49]  Norman E. Hawk,et al.  Steady and Unsteady Motions and Wakes of Freely Falling Disks , 1964 .

[50]  Hod Lipson,et al.  Unshackling evolution , 2014 .

[51]  S. Brunton,et al.  Discovering governing equations from data by sparse identification of nonlinear dynamical systems , 2015, Proceedings of the National Academy of Sciences.

[52]  Hod Lipson,et al.  Automated reverse engineering of nonlinear dynamical systems , 2007, Proceedings of the National Academy of Sciences.

[53]  Z. J. Wang,et al.  Falling paper: Navier-Stokes solutions, model of fluid forces, and center of mass elevation. , 2004, Physical review letters.