Soft robotics is a subfield of robotics that concerns the design, control, and fabrication of robots composed of compliant materials, instead of rigid links. In contrast to rigid-bodied robots built from metals, ceramics and hard plastics, the compliance of soft robots can improve their safety when working in close contact with humans.
Types and designs
The goal of soft robotics is the design and construction of robots with physically flexible bodies and electronics. In some applications, softness is restricted to a localized region of a machine. For example, rigid-bodied robotic arms can employ soft end effectors to gently grab and manipulate delicate or irregularly shaped objects. Most rigid-bodied mobile robots also strategically employ soft components, such as foot pads to absorb shock or springy joints to store/release elastic energy. However, the field of soft robotics generally focuses on the creation of machines that are predominately or entirely soft. Robots with entirely soft bodies have tremendous potential such as flexibility which allows them to squeeze into places rigid bodies cannot, which could prove useful in disaster relief scenarios. Soft robots are also safer for human interaction and for internal deployment inside a human body.
Nature is often a source of inspiration for soft robot design given that animals themselves are mostly composed of soft components and they appear to exploit their softness for efficient movement in complex environments almost everywhere on Earth. Thus, soft robots are often designed to look like familiar creatures, especially entirely soft organisms like octopuses. However, it is extremely difficult to manually design and control soft robots given their low mechanical impedance. The very thing that makes soft robots beneficial—their flexibility and compliance—makes them difficult to control. The mathematics developed over the past centuries for designing rigid bodies generally fail to extend to soft robots. Thus, soft robots are commonly designed in part with the help of automated design tools, such as evolutionary algorithms, which enable a soft robot's shape, material properties, and controller to all be simultaneously and automatically designed and optimized together for a given task.
Bio-mimicry Plant cells can inherently produce hydrostatic pressure due to a solute concentration gradient between the cytoplasm and external surroundings (osmotic potential). Further, plants can adjust this concentration through the movement of ions across the cell membrane. This then changes the shape and volume of the plant as it responds to this change in hydrostatic pressure. This pressure derived shape evolution is desirable for soft robotics and can be emulated to create pressure adaptive materials through the use of fluid flow. The following equation models the cell volume change rate:
is the rate of volume change. is area of the cell membrane. is the hydraulic conductivity of the material. is the change in hydrostatic pressure. is the change in osmotic potential.
This principle has been leveraged in the creation of pressure systems for soft robotics. These systems are composed of soft resins and contain multiple fluid sacs with semi-permeable membranes. The semi-permeability allows for fluid transport that then leads to pressure generation. This combination of fluid transport and pressure generation then leads to shape and volume change.
Another biologically inherent shape changing mechanism is that of hygroscopic shape change. In this mechanism, plant cells react to changes in humidity. When the surrounding atmosphere has a high humidity, the plant cells swell, but when the surrounding atmosphere has a low humidity, the plant cells shrink. This volume change has been observed in pollen grains and pine cone scales.
Similar approaches to hydraulic soft joints can also be derived from arachnid locomotion, where strong and precise control over a joint can be primarily controlled through compressed hemolymph.
Manufacturing Conventional manufacturing techniques, such as subtractive techniques like drilling and milling, are unhelpful when it comes to constructing soft robots as these robots have complex shapes with deformable bodies. Therefore, more advanced manufacturing techniques have been developed. Those include Shape Deposition Manufacturing (SDM), the Smart Composite Microstructure (SCM) process, and 3D multi-material printing.
SDM is a type of rapid prototyping whereby deposition and machining occur cyclically. Essentially, one deposits a material, machines it, embeds a desired structure, deposits a support for said structure, and then further machines the product to a final shape that includes the deposited material and the embedded part. Embedded hardware includes circuits, sensors, and actuators, and scientists have successfully embedded controls inside of polymeric materials to create soft robots, such as the Stickybot and the iSprawl.
SCM is a process whereby one combines rigid bodies of carbon fiber reinforced polymer (CFRP) with flexible polymer ligaments. The flexible polymer acts as joints for the skeleton. With this process, an integrated structure of the CFRP and polymer ligaments is created through the use of laser machining followed by lamination. This SCM process is utilized in the production of mesoscale robots as the polymer connectors serve as low friction alternatives to pin joints.
Additive manufacturing processes such as 3D printing can now be used to print a wide range of silicone inks using techniques such as direct ink writing (DIW, also known as Robocasting). This manufacturing route allows for a seamless production of fluidic elastomer actuators with locally defined mechanical properties. It further enables a digital fabrication of pneumatic silicone actuators exhibiting programmable bioinspired architectures and motions. A wide range of fully functional soft robots have been printed using this method including bending, twisting, grabbing and contracting motion. This technique avoids some of the drawbacks of conventional manufacturing routes such as delamination between glued parts. Another additive manufacturing method produces shape morphing materials whose shape is photosensitive, thermally activated, or water responsive. Essentially, these polymers can automatically change shape upon interaction with water, light, or heat. One such example of a shape morphing material was created through the use of light reactive ink-jet printing onto a polystyrene target.
Additionally, shape memory polymers have been rapid prototyped that comprise two different components: a skeleton and a hinge material. Upon printing, the material is heated to a temperature higher than the glass transition temperature of the hinge material. This allows for deformation of the hinge material, while not affecting the skeleton material. Further, this polymer can be continually reformed through heating.
Control methods and materials All soft robots require an actuation system to generate reaction forces, to allow for movement and interaction with its environment. Due to the compliant nature of these robots, soft actuation systems must be able to move without the use of rigid materials that would act as the bones in organisms, or the metal frame that is common in rigid robots. For actuation that involves bending, some sort of stress difference must be created across the component, such that the system has a tendency to bend towards a certain shape to relieve this said stress. Nevertheless, several control solutions to soft actuation problem exist and have found its use, each possessing advantages and disadvantages. Some examples of control methods and the appropriate materials are listed below.
Electric field One example is utilization of electrostatic force that can be applied in:
Dielectric Elastomer Actuators (DEAs) that use high-voltage electric field in order to change its shape (example of working DEA). These actuators can produce high forces, have high specific power (W kg−1), produce large strains (>1000%), possess high energy density (>3 MJ m−3), exhibit self-sensing, and achieve fast actuation rates (10 ms - 1 s). However, the need for high-voltages quickly becomes the limiting factor in the potential practical applications. Additionally, these systems often exhibit leakage currents, tend to have electrical breakdowns (dielectric failure follows Weibull statistics therefore the probability increases with increased electrode area ), and require pre-strain for the greatest deformation. Some of the new research shows that there are ways of overcoming some of these disadvantages, as shown e.g. in Peano-HASEL actuators, which incorporate liquid dielectrics and thin shell components. These approach lowers the applied voltage needed, as well as allows for self-healing during electrical breakdown.
Thermal
Shape memory polymers (SMPs) are smart and reconfigurable materials that serve as an excellent example of thermal actuators that can be used for actuation. These materials will "remember" their original shape and will revert to it upon temperature increase. For example, crosslinked polymers can be strained at temperatures above their glass-transition (Tg) or melting-transition (Tm) and then cooled down. When the temperature is increased again, the strain will be released and materials shape will be changed back to the original. This of course suggests that there is only one irreversible movement, but there have been materials demonstrated to have up to 5 temporary shapes. One of the simplest and best known examples of shape memory polymers is a toy called Shrinky Dinks that is made of pre-stretched polystyrene (PS) sheet which can be used to cut out shapes that will shrink significantly when heated. Actuators produced using these materials can achieve strains up to 1000% and have demonstrated a broad range of energy density between <50 kJ m−3 and up to 2 MJ m−3. Definite downsides of SMPs include their slow response (>10 s) and typically low force generated. Examples of SMPs include polyurethane (PU), polyethylene teraphtalate (PET), polyethyleneoxide (PEO) and others. Shape memory alloys are behind another control system for soft robotic actuation. Although made of metal, a traditionally rigid material, the springs are made from very thin wires and are just as compliant as other soft materials. These springs have a very high force-to-mass ratio, but stretch through the application of heat, which is inefficient energy-wise.
Pressure difference
Pneumatic artificial muscles, another control method used in soft robots, relies on changing the pressure inside a flexible tube. This way it will act as a muscle, contracting and extending, thus applying force to what it's attached to. Through the use of valves, the robot may maintain a given shape using these muscles with no additional energy input. However, this method generally requires an external source of compressed air to function. Proportional Integral Derivative (PID) controller is the most commonly used algorithm for pneumatic muscles. The dynamic response of pneumatic muscles can be modulated by tuning the parameters of the PID controller.
Sensors
Sensors are one of the most important component of robots. Without surprise, soft robots ideally use soft sensors. Soft sensors can usually measure deformation, thus inferring about the robot's position or stiffness.
Here are a few examples of soft sensors:
Soft stretch sensors Soft bending sensors Soft pressure sensors Soft force sensors
These sensors rely on measures of:
Piezoresistivity: polymer filled with conductive particles,Stassi, Stefano, et al. "Flexible tactile sensing based on piezoresistive composites: A review." Sensors 14.3 (2014): 5296-5332.