PETER: A Soft Pneumatic Manipulator with High Resistance and Load Capacity

Abstract

Despite the numerous advantages offered by pneumatic actuators, it remains challenging to design robots that can simultaneously meet the diverse demands typically taken for granted in traditional robotics, such as repeatability, precision, and load capacity. This work presents PETER, a robotic platform formed by three actuators with high extension capacity and which, in addition, exhibits high repeatability, notable resistance to fatigue and, most notably, a high load capacity. This is made possible by the utilisation of a TPU structure deformed as a consequence of the inflation rubber airbags. To evaluate these attributes, a closed-loop controller based on neural networks has been implemented on PETER. Despite its simplicity, this controller achieves a precision of 2°and is able to carry weights up to 500 g. The potential avenues for further research and development are outlined.