Our research in soft-material robotics differs from earlier robotic approaches focused on systems (both animals and robots) with rigid joints and centralized control systems that use discrete commands. Our research addresses the biggest challenge in soft robotics: controlling soft materials. Soft materials are difficult to control because of intrinsic non-linear behavior.

Contrast a rigid rod with an elastic rubber, representing soft tissue. Given the initial position of a rod we can easily predict its kinematics (motion) when a local force is applied and under most conditions this is highly repeatable. Now consider an elastic rubber: its kinematics in response to a local force is much harder to predict and the way the force propagates though the material is complex. Furthermore, the repeated application of a force might have different consequences because soft materials show marked history-dependency (such as non-linear pseudo-elasticity and work softening). Clearly, controlling soft materials to induce locomotion requires the application and timing of a set of forces entirely different than those required to move a rigid rod. The complexity is increased tremendously for a system composed of elastic elements capable of large deformations.

One component of our research is investigating how soft-bodied animals control their locomotion through neuromechanics (tight coupling of neural and mechanical processes).


Manduca sexta larvae,
Taken from: http://en.wikipedia.org/wiki/Image:Tobacco_Hornworm.png

To investigate one particular aspect of controlling locomotion in a soft-bodied animal, Manduca sexta caterpillar: Given a Manduca model, we develop a computational solution that elicits locomotion in the model. The solution consists of applying forces to different Manduca muscles at certain intervals. The solution represents firing of the central nervous system. The soft-bodied caterpillar Manduca sexta is used as our model organism because it offers a simple platform for analyzing how interactions between neural and biomechanical elements control locomotion.