Dynamic Morphological Computation: Embodying Control Intelligence in Soft Material Robots

Having to interact with people or dynamic environments, modern robots cannot rely anymore on rigid links governed by a centralized control only. In order to make a system adaptable, some sort of intelligence needs to be embedded directly in the body of the robot itself. A key element to achieve such feature, also known as morphological computation, is the use of soft materials. In fact, in biological systems, softness is a fundamental property used to respond to external and internal forces. Taking inspiration from nature, soft robotics aims at enhancing robotic performances through the employment of soft materials. The present thesis fits into this new emerging field. Most of the state-of-the-art soft robots are fluidic systems, composed of several chambers, independently controlled. Typically, this design requires a complex and bulky network of active components for actuation and control. Starting from this observation, a novel approach to fluidic under-actuation is investigated, focusing on the dynamical response of the system. This idea draws inspiration from a standard technique applied in analog electronics and mechanical design, which consists in tuning the frequency response of systems. The proposed method is based on the co-design both of the mechanical parameters of the system and of custom input signals, to enable the elicitation of different behaviors with fewer control components. The suggested principle is presented in theory and simulation, and then experimentally validated through the application to a case study, an in-pipe inchworm-like robot. It is shown that it is possible to obtain forward and backward movements by modulating a unique pressure input. In this first example, stiffness and damping of the chambers are regulated through separate dashpots and elastic bands. In the next step we implement integrated soft chamber systems with desired mechanical properties. For the involved soft materials (silicone in our case), the main challenge consists in independently changing the properties of stiffness and damping. Several approaches have been analyzed via simulations and experiments and the most effective solution has been employed to realize different types of soft actuators. In particular, working principle and fabrication process of extending, contracting and bending actuators are described. Finally, an application is shown, involving a two-fingered gripper fed by a single pneumatic line, which is able to perform pinch and power grasp motions.