Soft Robotics Grippers
By Shashank Rajiv Moghe
Soft robotics was a course I took taught by Dr. Jeane D`Souza in my seventh semester of my college. The course itself proved to be mostly theory oriented, and the reason for that became apparent very soon. It’s a completely foreign paradigm to someone studying conventional robotics - working with rigid structures, deterministic frames and kinematics around those - soft robotics seems like an weird pseudoscience, requiring feedback control and real world testing even for prototypes.
There were very few simulators that would help simulate these deformable systems, and a precious few which were accessible to college students. My requirements for a simulator (or any software for that matter) for my personal use are that they should be (a) Free and Open Source and (b) have a reasonably supportive community. MathWorks’ MatLAB had SoRoSim, which I would rather not use because of the proprietary nature (I didn’t check if it worked with GNU Octave though). COMSOL and DassaultSystemes’ Abaqus, both which were recommended to me by some professors, were no good for me, as they were not only proprietary, but expensive too.
This led me to the SOFA-Framework which fit my requirements perfectly. I spent a few weeks learning SOFA and its quirks, and copied the basic implementation of a pneunet. I was able to simulate a simple, well functioning (and most importantly stable) pneunet which could show you the reaction to different pressures applied to the deformable system. You can checkout the docker container I used here.
With this implemented I asked my professor for any quick projects I could work on to understand the system better, and she gave gave me one.
I worked on setting up the basics of a soft gripper using LSOVAs, and then worked on integrating more components into this to make it a gripper. This turned into a long and ongoing topic of research at the college, led by more interested students now. I however can share the first implementation of LSOVAs in SOFA here.
While in total this is not much ‘work’, I learnt a lot and started a new topic of research, and for that, I’m quite happy. I am intrigued greatly by soft robotics as a concept, and am looking to work on this more in the future.