Meet the octobot, the first robot to be made entirely from soft materials. Powered by a chemical reaction and controlled by a fluidic logic circuit, it heralds a generation of soft robots that might surpass conventional machines. See Letter p.451
Robots are typically used in manufacturing contexts that involve well-structured environments. These situations allow them to move following predefined procedures, limiting interactions with human operators for safety reasons. But if these machines were moved into 'real' environments outside factories, they would have to cope with uncertain situations, react and adapt to changing conditions, and interact safely with living organisms, including humans1 — tough problems to solve using conventional technology made from hard materials. Robots made from soft, deformable materials2 would be better able to grasp and manipulate unknown objects, and to move on unstructured and rough terrains, and might be less hazardous to people. On page 451, Wehner et al.3 present the first robot that completely lacks rigid structures and control systems.
Soft body parts are important in many natural organisms. Animals such as squid, starfish and worms are composed almost entirely of soft materials and liquids, which increases their adaptability and robustness. There is therefore a growing belief that soft materials might help robotics technology to go beyond its current capabilities, by allowing robots to elongate, squeeze, climb and grow. For example, soft robotic arms inspired by octopuses can elongate4, and soft robots that mimic caterpillars can roll and jump5.
A notable attempt to develop a fully soft robot was reported6 in 2011 by workers from the same research group as that of Wehner and colleagues. In that case, the robot itself was composed exclusively of soft materials, but a conventional pump-and-valve system was used to implement (actuate) different types of locomotion pneumatically, and was connected to the robot through cables. Wehner and colleagues now push the technological boundaries further, because not only are their robot's body and actuation units soft, but so also are the control system and power source, which are integrated into the robot. This makes it the first completely soft robot capable of operating without being tethered by cables.
The octopus-shaped robot — dubbed the 'octobot' by the authors (Fig. 1) — has eight arms moved by a pneumatic mechanism that relies on the expansion of embedded, inflatable compartments working as actuators. These actuators are integrated into a fluidic–pneumatic network powered by a liquid fuel (an aqueous solution of hydrogen peroxide). The fuel passes through reaction chambers that contain a platinum-based catalyst, which causes the hydrogen peroxide to decompose. This decomposition produces pressurized oxygen that inflates the actuators, thus generating the arm movements.
Wehner et al. control the sequence of the octobot's arm movements using a completely soft fluidic circuit based on a system of valves that act as elements of logic gates. The circuit creates an oscillation that converts the inflow of pressurized fuel from the fuel storage chamber into outflows that alternate between different reaction chambers, until the system runs out of fuel. The octobot therefore repeats cycles of movements in which it first lifts four of its arms while lowering the other four, and then performs the reverse manoeuvre (see go.nature.com/2b3cn3s). The whole of the robot's body, including the fluidic circuit, is made of silicone-based materials that have different mechanical properties, tailored to the functional requirements of the various subsystems.
The realization of autonomous soft robots will require the integration of different materials and functionalities, such as actuation, powering and logic; the octobot represents the minimal system that demonstrates the potential of this approach. To achieve the required integration, Wehner and colleagues used a combination of advanced fabrication techniques — including micro-moulding7, soft lithography8 and multi-material embedded 3D printing9 — to produce rubbery structures embedded with fluidic channels, spanning several orders of magnitude in length scales. Despite its apparent complexity, the customizability of this fabrication process allowed the authors to validate design modifications using a quick trial-and-error approach, so that the final device was rapidly optimized.
Wehner and colleagues' use of soft materials and continuum deformations — continuous bending of the arms to generate movement, rather than motion created by rigid structures connected by rotational joints — paves the way for further scientific and technological developments. The next steps are to develop computational control systems (such as more-sophisticated fluidic circuits) that allow a greater range of movement; to define new design rules for soft robots; and to adopt and improve manufacturing technologies.
Other challenges remain. For example, the forces that soft robots can exert on the environment might be limited, potentially restricting their applications. Moreover, the use of fluidic logic circuits as control systems, rather than conventional electronics, might limit the complexity of the behaviours that can be generated. A greater understanding of the properties of soft materials and how they interact with control systems and the environment is also needed, to produce desired robotic behaviour in real contexts10.
Although soft robotics is still in its infancy, it holds great promise for several applications, such as servicing and inspecting machinery, search-and-rescue operations, and exploration. Soft robots might also open up new approaches to improving wellness and quality of life. Soft endoscopes that allow omnidirectional bending, elongation and tunable stiffening are already a reality11, as are soft orthotic devices used for ankle and foot rehabilitation12. Wehner and colleagues' findings might help to guide research in these directions, contributing to the pillars of knowledge that will support the edifice of this new discipline.Footnote 1
Pfeifer, R., Lungarella, M. & Iida, F. Science 318, 1088–1093 (2007).
Rus, D. & Tolley, M. T. Nature 521, 467–475 (2015).
Wehner, M. et al. Nature 536, 451–455 (2016).
Mazzolai, B., Margheri, L., Cianchetti, M., Dario, P. & Lasch, C. Bioinspir. Biomim. 7, 025005 (2012).
Lin, H.-T., Leisk, G. G. & Trimmer, B. Bioinspir. Biomim. 6, 026007 (2011).
Shepherd, R. F. et al. Proc. Natl Acad. Sci. USA 108, 20400–20403 (2011).
Heckele, M. & Schomburg, W. K. J. Micromech. Microeng. 14, R1–R14 (2004)
Qin, D., Xia, Y. & Whitesides, G. M. Nature Protocols 5, 491–502 (2010).
Wu, W., DeConinck, A. & Lewis, J. A. Adv. Mater. 23, H178–H183 (2011).
Paul, C. Robot. Auton. Syst. 54, 619–630 (2006).
Ranzani, T., Gerboni, G., Cianchetti, M. & Menciassi, A. Bioinspir. Biomim. 10, 035008 (2015).
Park, Y.-L. et al. Bioinspir. Biomim. 9, 016007 (2014).
About this article
Nature Communications (2019)
Nature Reviews Materials (2019)
Journal of Materials Science (2019)
Current Otorhinolaryngology Reports (2017)