Inverse kinematics of soft robots in their environment

Soft robotics is a very recent and active field where researchers are actively exploring robots designs and their usages. Soft robotics draws its inspiration from nature, from the way living organisms move and adapt their shape to their environment. The use of soft materials allow these robots to accomplish tasks with more flexibility and adaptability.

Solving the inverse kinematics of a soft robot can not be achieved directly from standard methods used in traditional robotics. The main reason is related to the motion of soft robots obtained through deformation of the structure rather than by articulations. Therefore, the behavior of soft robots should be modeled using deformable mechanics. The use of continuum mechanics raises several issues. No analytic solution exists in the general case and numerical methods, typically the finite element method, have to be used.

In addition, due to their natural compliance, soft robots are often used in contact with their environment. Yet, their kinematics is highly dependent on external factors, which increases the complexity of solving their inverse kinematics.

In this document we detail methods we propose for the modeling of a wild range of soft robots, and methods for solving their inverse kinematics. The methods are based on the finite element method to capture the deformations of the robot’s structure, and of its environment when deformable. We formulate the problem of their inverse kinematics as an optimization program, allowing easy handling of constraints on actuation and singularity problems. These methods we propose also include contacts into the optimization process.

We give a particular attention to provide solutions with real-time performance, allowing online control in evolving environments.