What is a soft robot ?

An illustration of a typical industrial robot made using the midjourney AI.

What is a soft robot in the first place ?

A soft robot is a robot whose overall structure is flexible or compliant. This means it can deform when pushed or pulled. 

To reach compliance, various strategies can be used. One of the most straightforward strategy is to use soft materials. In scientific literature, a material's softness is usually measured using its Young's modulus represented by the letter E and measured in Pascals. This is the ratio of the amount of pressure applied on a given material to it's deformation: E = σ / ε

A high Young's modulus means a material is hard, while soft materials have a low Young's modulus. Aluminum has a Young's modulus of about 63 GPa, or 63,000,000,000 Pascals.  Materials used in soft robotics such as silicone elastomers can have Young's modulus as low as 100 kPa, or 100,000 Pascals. That's about  630,000 times softer. 

For example, Baymax,  the robot from the animated movie Big Hero 6 from Pixar, was inspired by real life research from Prof. Chris Atkeson at Carnegie-Mellon University [1].  In the movie, the robot is made of a flexible membrane that is inflated

 Another strategy to create a compliant structure is to assemble rigid materials in ways that allow them to deform. For example a knitted piece of fabric is quite comfortable to wear. This is because, although each individual thread is fairly rigid, their assembly makes the whole piece of fabric quite stretchable.   

An illustration of a soft inflatable robot arm made using the midjourney AI.

How do soft robots differ from "hard" traditional robots ?

Materials. Soft robots their traditional counterparts are not designed using the same materials. Soft robots are made of soft and/or flexible materials, while traditional robots are made of mostly rigid materials. From this difference stems the need to design and control the robots differently, as well as the use of the robots in distinct applications. 

Applications. Traditional robots usually operate highly controlled environments and do highly repetitive tasks. They usually operate within the scope of the "three Ds":  they replace human workers in jobs that are either Dirty, Dull, or Dangerous. Traditional robots can be dangerous pieces of machinery, and in fact, in factories humans usually don't work within their range. 

Soft robots aim to solve problems that traditional robots can't and won't be able to solve. They do not aim at replacing traditional robotics. Rather, the goal is to add robots in new environments to solve problems that traditional robots can't. 

Medical environments, where softness is important, or search and rescue missions, where adaptability is key, may be good future applications of soft robotics [2,3]. Food and agricultural applications may also benefit from soft robotics due to improved grasping softness and adaptability.  

How's the field of soft robotics emerging ?

The field of soft robotics is, for now, mostly subject to academic research. There are few industrial applications of soft robotics at scale. The growth of the academic research field can somewhat be measured in the number of peer-reviewed scientific publication each year. Such an analysis shows a rapid increase in publication from 2010 to recent years [4], with industrial endeavors having also met some success [5,6]. 

For the sake of simplicity, so far, we have distinguished between "traditional" and "soft" robots. A more accurate depiction would be that robots belong on a  spectrum. As an example, consider Collaborative Robots. CoBots are robots made using traditional techniques and materials but rated to share the work environment with humans [7]. On a scale, one could argue that they belong somewhere between traditional factory robots on one end, and soft robots on then other end. 

From the point of view of a robot user, soft robotics should be the solution to problems that traditional robotics can't solve. Because soft robotics is relatively new, it's unlikely that we will see entirely soft robots used outside the laboratories anytime soon, as many technical challenges have yet to be solved. However, what we could see happen is the design of soft appendages mounted on more rigid robotic platforms with the goal of solving previously intractable problems. Rather than expecting to see soft robots in real life soon, one could envision robotic platforms mounted with soft hands, tentacles, legs or skin and thus benefiting from increased resilience, adaptability, and capability.

References for further reading: 

[1] https://globalnews.ca/news/1663166/big-hero-6-star-baymax-inspired-by-robotics-research/, accessed 28/06/2023

 [2] Trimmer, Barry. “A Journal of Soft Robotics: Why Now?” Soft Robotics 1, no. 1 (March 2014): 1–4. https://doi.org/10.1089/soro.2013.0003.

[3] Majidi, Carmel. “Soft Robotics: A Perspective—Current Trends and Prospects for the Future.” Soft Robotics 1, no. 1 (March 2014): 5–11. https://doi.org/10.1089/soro.2013.0001.

[4] Hawkes, Elliot W., Carmel Majidi, and Michael T. Tolley. “Hard Questions for Soft Robotics.” Science Robotics 6, no. 53 (April 28, 2021): eabg6049. https://doi.org/10.1126/scirobotics.abg6049.

[5] Amend, John, Nadia Cheng, Sami Fakhouri, and Bill Culley. “Soft Robotics Commercialization: Jamming Grippers from Research to Product.” Soft Robotics 3, no. 4 (December 2016): 213–22. https://doi.org/10.1089/soro.2016.0021.

[6] https://www.softroboticsinc.com/, accessed 31/07/2023

[7] Wikipedia contributors, "Cobot," Wikipedia, The Free Encyclopedia, https://en.wikipedia.org/w/index.php?title=Cobot&oldid=1172607955 (accessed September 12, 2023).