Frontiers in Robotics and AI (Aug 2021)
Hardware Methods for Onboard Control of Fluidically Actuated Soft Robots
Abstract
Soft robots provide significant advantages over their rigid counterparts. These compliant, dexterous devices can navigate delicate environments with ease without damage to themselves or their surroundings. With many degrees of freedom, a single soft robotic actuator can achieve configurations that would be very challenging to obtain when using a rigid linkage. Because of these qualities, soft robots are well suited for human interaction. While there are many types of soft robot actuation, the most common type is fluidic actuation, where a pressurized fluid is used to inflate the device, causing bending or some other deformation. This affords advantages with regards to size, ease of manufacturing, and power delivery, but can pose issues when it comes to controlling the robot. Any device capable of complex tasks such as navigation requires multiple actuators working together. Traditionally, these have each required their own mechanism outside of the robot to control the pressure within. Beyond the limitations on autonomy that such a benchtop controller induces, the tether of tubing connecting the robot to its controller can increase stiffness, reduce reaction speed, and hinder miniaturization. Recently, a variety of techniques have been used to integrate control hardware into soft fluidic robots. These methods are varied and draw from disciplines including microfluidics, digital logic, and material science. In this review paper, we discuss the state of the art of onboard control hardware for soft fluidic robots with an emphasis on novel valve designs, including an overview of the prevailing techniques, how they differ, and how they compare to each other. We also define metrics to guide our comparison and discussion. Since the uses for soft robots can be so varied, the control system for one robot may very likely be inappropriate for use in another. We therefore wish to give an appreciation for the breadth of options available to soft roboticists today.
Keywords