Large-scale soft robots for reliable manipulation

Traditional robots are limited in their reach unless they become large and heavy. We wish to develop large-scale soft robots with 7 or more degrees of freedom that have high pressure air-filled bladders reinforced with nylon fabric for structural purposes, and pleated variable-pressure bladders for joints. These robots will be able to reach long distances with very little weight. However, the behavior of these air-filled chambers is complex and their deformation may not follow predictions due to changes in load in pose. We are seeking funding to rediscover a reliable manufacturing process for these nylon-reinforced soft robots, integrate modern sensors that give reliable information about their motions and state of `health' (e.g is a collapse or crease immanent) and create more precise models of their behavior. We will develop estimation methods that are tailored to the specific ways in which soft robots behave, analyzing pleats, folds, textile mechanics, and gas dynamics. During the course of the project, we will begin by creating modules for single degree of freedom joints, and develop plug-ins for the disc-thread model of soft robots that we developed in our prior work. These modules will be sensorized using our optical fiber network sensors, also developed as part of our prior work. We will advance a high-fidelity simulation environment for these modules with dynamic equations that capture the subtle mechanical interactions of the bladder and the fabric, and validate the simulation using experimental data derived from our modules. Having done this, we will pursue more integrated designs of the previously developed modules to solve the problem of creating a realistic, functional six-degree-of-freedom soft robot with a reach of over 2 meters. The final result will be a soft robot prototype that can lift over 10 times its weight.