Light activated soft robotics- shankar research group

The Shankar research group at the University of Pittsburgh is headed by professor Ravi M. Shankar and studies Liquid Crystalline Elastomers. Materials that due to their structural asymmetry, deform when exposed to beams of light. The lab focused on developing robots that could use this material for actuation, creating robots that required absolutely no electronics, no wires, and no tether. Soft robotics is a very new area of robotics research so much of the research consisted of the study of the material and its properties in different environments.

Upon beginning work with the lab, a small crawling microrobot had been constructed and tested, yielding promising results. The next step was the creation of a swimming robot, that used a strip of liquid crystal elastomer as the tail. by oscillating this strip one could create the tail movement necessary for swimming

We began by testing the materials actuation power underwater. We also studied light attenuation, which would affect the intensity of light beams shown on materials underwater. Test results were promising, but the Univerisity of Pittsburgh shifting all work to remote in march required us to take a different approach to study the material. Under the supervision of professor Kaushik Dayal of Carnegie Mellon University, I began to program a two-dimensional simulation that modeled the behavior of a strip of liquid crystal elastomer. 

The simulation would also provide information about the internal properties of the material such as stress, strain, and ionized molecule concentration along the beam over time. These simulations were based on a set of partial differential equations from a previous research paper, “A Non-Linear Beam Model of Photomotile Structures” By Kevin Korner. I added in a flight control system that would allow the user to actuate the beam in different ways with different lighting patterns

The other aspect of the simulation I had to consider was the properties of the strip underwater. In order to properly model the interaction of the beam with water, I studied Purcell’s scallop theorem which discusses movement in water for incredibly small objects. I also attempted to add drag force to the beam due to water, but the net linear drag force from each oscillation ended up being negligible.

Purcell’s theorem states that non-reciprocal motion is impossible in low Reynolds number environments if the object only has one degree of freedom. To remedy this I attached a second LCE strip to the first one and used a different oscillation pattern to create cyclic motion, allowing for forward swimming. 

In order to add the water physics to my simulation, I had to program the Navier stokes equations into Matlab. I used a combination of reference code found on youtube as well as information from NVIDIA’s website to create my program. The result was a tank around my beam with water represented by velocity vectors. Based on the location of the beam over time, I created a solid mesh to interact with the Navier Stokes equations allowing the movement of the beam to influence the water in real-time. By integrating the velocity water field I could determine the net displacement of the water and therefore the net displacement that a robot would incur as a result of a specific oscillation pattern. This was ultimately what I set out to do, as I could now determine how a specific length and oscillation pattern, and light intensity would affect the movement and force on the robot.

The last thing I did was perform fluid simulations with various Solidworks bodies to determine which microrobot bodies had the best potential. I did this by setting the speed of the water equal to the speed of the robot calculated in our Matlab simulation and determining the drag coefficient of each body. Doing this provided a better idea of which type of swimmer bodies would be best.

The final result was a comprehensive Liquid crystal elastomer simulation underwater that could be used to test different material properties and movement patterns for a theoretical swimming microrobot. By using this tool, we can study our robot and material while remote so that when we finally get back in the lab, we can construct the robot more easily and effectively.