Matt Travers and Howie Choset
Carnegie Mellon University
Our attempts to mimic animal motion have resulted in many technological advances that have revolutionized how manmade machines move through air, in water, and over land. Despite numerous achievements, engineers and scientists have yet to closely replicate the grace and fluidity of animal movement. This suggests the biological world still has much in the way of suggestions for how to build, design, and program robotic systems whose locomotive capabilities will far outpace what is possible today. The question then becomes: How deeply should we look at biology? Take the transition from snake to snake robot as an example. On the surface, one can see a snake, say, on a hike in the woods and then build an elongated mechanical creature. However, we can go deeper: One can study the fundamental macroscopic principles that can be transferred from muscles and skeleton to conventional motors and mechanical linkages. Going even deeper, one can try to create new muscle-like actuators and controllers based on neural networks in an attempt to accurately copy biological function and control. The right choice of where to focus on this spectrum remains an open question.
To help address these fundamental questions, the biologically inspired robotics community has to date produced many great works, far too many to summarize in one brief article. Instead, we focus the attention of this short comment on what works have specifically inspired our research in the Biorobotics lab at Carnegie Mellon University over the past 20 years. In this time, we have built a number of different robots but are perhaps best known for novel snake-like robots (see http://biorobotics.org).
In our opinion, the single biggest influence in biological inspiration is Bob Full. His group at Berkeley studies cockroaches, crabs, and geckos, just to name a few. Full’s research interest is primarily in comparative biomechanics and physiology [1, 2]. He collaborates with a number of different engineers and other scientists to elucidate biological principles that inspire the design of advanced robotic components, control algorithms, and novel system designs.
Full’s work on geckos lead to a fundamental understanding of how their feet stick to nearly any surface and yet not are not encumbered by dirt and other particles. His collaborator Ron Fearing, also at Berkeley, developed new MEMS manufacturing technology to replicate the capabilities of geckos’ feet. Fearing’s work harnesses features of animal manipulation, locomotion, sensing, actuation, mechanics, dynamics, and control strategies to radically improve robotic capabilities, especially at very small scales. Fearing’s research ranges from the fundamental understanding of mechanical principles to novel fabrication techniques and system integration for autonomous machines [3, 4].
At Harvard University, Rob Wood also develops novel robotic mechanisms at very small scales [5, 6]. His work uses microfabrication techniques to develop biologically inspired robots with features on the micrometer to centimeter scales. His specific interests include new micro- and mesoscale manufacturing techniques, fluid mechanics of flapping wings, control of sensor-limited and computation-limited systems, active soft materials, and morphable soft-bodied robots.
In addition to novel designs and methods for constructing robot morphologies, biology also inspires us to design improved software to enable robots to better interact with complex environments. Shigeo Hirose is one of the early pioneers in the creation of numerous biologically inspired robotic systems that specialize in weaving their way through complex terrains. He is probably best known for his original work on serpentine locomotion, both analyzing the fundamental physics governing how biological snakes move as well as employing the lessons learned therein to create and control numerous mechanisms over the years . His ground-breaking insight into biologically inspired control has naturally influenced his robot’s designs.
Dan Koditschek’s name is synonymous with robot control, especially in the area of dynamic legged locomotion [8, 9]. He has played a major role in several seminal works in the area of bioinspired robots throughout his career (many in collaboration with Bob Full). In addition, he has overseen the construction of biologically inspired robots that have helped roboticists better understand mechanized locomotion as well as offered biologists better insight into the natural world. Additionally, Full’s observations on the role of compliance in mechanisms and control inspired Koditschek’s group to develop the family of RHex robots. Moreover, Koditschek and Full developed the concept of templates and anchors, a now-ubiquitous method for abstracting the motion of complex systems. Koditscheck’s more recent work with Full has started to explore both the design and the control of aerial acrobatics using tail-like appendages, originally inspired by the observation of geckos that control the orientation of their bodies using their tails in midair.
Related to the work by Full and Koditschek, the incorporation of compliance in robot mechanisms and control design can also be attributed to Gill Pratt. Pratt, in part, developed a new paradigm for robotic actuation–the series elastic actuator–as well controllers that employ this technology . This work has directly affected and certainly inspired several generations of robotic devices with different morphologies that move by slithering, crawling, and walking.
A. E. Hosoi’s research covers a diverse set of topics, from fundamentals of materials science and fluid dynamics to the control and practical application of locomotion and manipulation systems [11, 12]. Two projects that are particularly relevant to the study of biologically inspired robots are those that consider the Roboclam and the Robosnail. Both systems were constructed using direct biological inspiration aimed at practical real-world applications.
George Lauder’s work on fish-like robots has resulted in a series of robotic test platforms that examine fin and body kinematic and hydrodynamic functions during locomotion . Robotic devices have a considerable advantage over studying live fish in the sense that a variety of programmable motions permit the careful investigation of the discrete components of naturally coupled movements.
The idea of using a robot to serve as a surrogate to study biology is also present in Daniel Goldman’s work that focuses on studying systems that locomote on granular media [14-17]. Goldman, faculty and director of the Crablab at Georgia Tech, has recently coined the term “robophysics,” which relates to the practice of using robots as the basis for modeling biological systems in extremely complex terrains. Goldman’s team uses robots to help study snakes, lizards, ants, and turtles, just to name a few. Goldman’s group specializes in the interaction of physical and biological systems with complex materials, like granular media. His group looks investigates how organisms like lizards, crabs, and cockroaches generate appropriate musculoskeletal dynamics to scurry rapidly over substrates like sand, bark, leaves, and grass.
Noah Cowan has also applied and made novel advances in the application of control theory to the study of sensorimotor control of animal movement [18, 19]. He and his collaborators study weakly electric fish as well as cockroach antennae. At Northwestern, Malcolm MacIver, one of Cowan’s and Lauder’s collaborators, pursues a research program in the mechanical and neural basis of animal behavior, particularly at the intersection of information harvesting and biomechanics .
Finally, A. Ijspeert’s work on biologically inspired robots focuses on the computational aspects of locomotion control, sensorimotor coordination, and learning in animals and in robots [21, 22]. His group is interested in using robots and numerical simulation to study the neural mechanisms underlying movement control and learning in animals.
In the Biorobotics lab at Carnegie Mellon, inspiration has also been drawn from the works of J. Ostrowski and S. D. Kelly that employ concepts form the field of geometric mechanics to the study of undulatory locomotion [23-26]. In their respective works, Ostrowski and Kelly perform mathematical modeling, analysis, simulation, and control of systems that exhibit nonlinear dynamics. Former CMU student Elie Shammas, now faculty at the American University of Beirut, took this early work and developed visualization tools that enable intuition to guide the design of gaits for idealized articulated systems. Ross Hatton, who succeeded Shammas, took this work to the next level, generating results at the interface of robotics and applied mechanics [27, 28]. Hatton, now faculty at Oregon State University, provided a wealth of analytic tools to study snake-like locomotion as well as other locomoting systems. Recently, Hatton began new work that looks at sensing and control in spiders. Adding to the work of Goldman’s and Choset’s previous students, Chaohui Gong has recently created a new approach that brings to bear all of the analytic tools, developed by one of Choset’s students, to study both natural as well as robotic systems that locomote in granular media. Gong’s demonstrations include snake robots locomoting on rocks, sandy inclines, and in tight spaces.
1. T. Libby et al., Tail assisted pitch control in lizards, robots and dinosaurs. Nature 481, 181 (2012).
2. R. J. Full, T. Kubow, J. Schmitt, P. Holmes, D. Koditschek. Quantifying dynamic stability and maneuverability in legged locomotion. Int. Comp. Biol. 42, 149 (2002).
3. C. Li et al., Terradynamically streamlined shapes in animals and robots enhance traversability through densely cluttered terrain. Bioinspir. Biomim. 10, 046003 (2015).
4. A. G. Gillies et al., Gecko toe and lamellar shear adhesion on macroscopic, engineered rough surfaces. J. Exp. Biol. 217, 283 (2014).
5. M. A. Graule et al., Perching and takeoff of a robotic insect on natural and artificial overhangs using switchable electrostatic adhesion. Science 352, 978 (2016).
6. J.-S. Koh et al., Jumping on water: Surface tension–dominated jumping of water striders and robotic insects. Science 349, 517 (2015).
7. S. Hirose, Biologically Inspired Robots: Snake-Like Locomotors and Manipulators (Oxford University Press, Oxford, 1993).
8. A. Altendorfer et al., RHex: A biologically inspired hexapod runner. J. Autonomous Robots 11, 207 (2002).
10. G. A. Pratt, M. M. Williamson, Series elastic actuators, in vol. 1 of IEEE International Conference on Intelligent Robots and Systems (1995), pp. 399–406.
11. A. G. Winter et al., Razor clam to RoboClam: Burrowing drag reduction mechanisms and their robotic adaptation. Bioinspir. Biomim. 9, 036009 (2014).
12. B. Chan, N. J. Balmforth, A. E. Hosoi, Building a better snail: Lubrication and gastropod locomotion. Phys. Fluids 17, 113101 (2005).
13. G. V. Lauder, E. J. Anderson, J. Tangorra, P. G. A. Madden, Fish biorobotics: Kinematics and hydrodynamics of self-propulsion. J. Exp. Biol. 210, 2767 (2007).
14. B. McInroe et al., Tail use improves soft substrate performance in models of early vertebrate land locomotors. Science 353, 154 (2016).
15. H. C. Astley et al., Modulation of orthogonal body waves enables high maneuverability in sidewinding locomotion, Proc. Natl. Acad. Sci. U.S.A. 112, 6200 (2015).
16. J. Aguilar et al., A review on locomotion robophysics: The study of movement at the intersection of robotics, soft matter and dynamical systems. Rep. Prog. Phys. 79, 110001 (2016).
17. Tingnan Zhang, Daniel I. Goldman, The effectiveness of resistive force theory in granular locomotion. Phys. Fluids 26, 101308 (2014).
18. J. M. Mongeau, A. Demir, J. Lee, N. J. Cowan, R. J. Full, Locomotion and mechanics mediated tactile sensing: Antenna reconfiguration simplifies control during high-speed navigation in cockroaches. J. Exp. Biol. 216, 4530 (2013).
19. S. Sefati et al., Mutually opposing forces during locomotion can eliminate the tradeoff between maneuverability and stability. Proc. Natl. Acad. Sci. U.S.A. 110, 18798 (2013).
20. Y. Bai, J. B. Snyder, M. A. Peshkin, M. A. MacIver, Finding and identifying simple objects underwater with active electrosense. Int. J. Robotics Res. 34, 1255 (2015).
21. K. Karakasiliotis et al., From cineradiography to biorobots: An approach for designing robots to emulate and study animal locomotion, in J. R. Soc. Interface, 13, 119, (2016).
22. A. Ijspeert. Biorobotics: Using robots to emulate and investigate agile animal locomotion. Science, 346, 196 (2014).
23. J Ostrowski, J. Burdick, Gait kinematics for a serpentine robot, in IEEE International Conference on Robotics and Automation (IEEE, 1996).
24. J Ostrowski, J Burdick, The geometric mechanics of undulatory robotic locomotion. Int. J. Robotics Res. 17, 683 (1998).
25. S. D. Kelly, H. Xiong, Self-propulsion of a free hydrofoil with localized discrete vortex shedding: analytical modeling and simulation. Theor. Comput. Fluid Dyn. 24, 45 (2010).
26. P. Tallapragada, S. D. Kelly, Dynamics and self-propulsion of a spherical body shedding coaxial vortex rings in an ideal fluid. Dynamics 18, 21 (2013).
27. H. Faraji, R. L. Hatton, Aiming and vaulting: Spider-inspired leaping for jumping robots, in Proceedings of the IEEE International Conference on Robotics and Automation (IEEE, 2016).
28. H. Faraji et al., Impulse redirection of a tethered projectile, in Proceedings of the ASME Dynamic Systems and Controls Conference (DSCC), (ASME, 2015).