Pneu-HTP based extrusion actuation in bio-inspired soft joint

Inspired by the joint structure and actuation mechanism of spider legs, a novel pneumatic soft joint actuator is designed, which achieves joint rotation by mutual compression of two hyperelastic sidewalls under inflation pressure. For this type of extrusion actuation, a pneumatic hyperelastic thin plate (Pneu-HTP) based actuation modeling method is proposed. The two actuating surfaces extruded mutually of the actuator are considered as Pneu-HTPs, and mathematical models for their parallel extrusion actuation and angular extrusion actuation are derived. The finite element analysis (FEA) simulations and experiments were also performed to evaluate the model accuracy of the Pneu-HTP extrusion actuation. The results for the parallel extrusion actuation show that the average relative error between the proposed model and the experiment is only 9.27%, and the goodness-of-fit is greater than 99%. For the angular extrusion actuation, the average relative error between the model and the experiment is 12.5%, and the goodness-of-fit is greater than 99%. The parallel extrusion actuating force and rotational extrusion actuating force of the Pneu-HTP are also highly consistent with the FEA simulation results, which provides a promising method for the accurate modeling of extrusion actuation in soft actuator.

Automated summary

This paper presents a novel pneumatic soft joint actuator inspired by the joint structure and actuation mechanism of spider legs. The actuator achieves joint rotation through the mutual compression of two hyperelastic sidewalls under inflation pressure. A pneumatic hyperelastic thin plate (Pneu-HTP) based actuation modeling method is proposed for this extrusion actuation, and mathematical models for parallel and angular extrusion actuation are derived. Finite element analysis simulations and experiments are conducted to evaluate the accuracy of the Pneu-HTP extrusion actuation model. The results show high consistency between the model and the experiments, with average relative errors ranging from 9.27% to 12.5% and goodness-of-fit greater than 99%.