Analysis and modeling of nonlinear saturated behavior of layer jamming soft pneumatic bending actuator

Soft actuators have become remarkably popular among numerous applications in rehabilitation and manipulation. Despite their numerous advantages, these actuators exhibit a significant limitation in grasping applications. Their inherently low stiffness, a characteristic of soft actuators, leads to considerable deformation when interacting with opposing forces. In this study, a jamming actuator has been integrated into the soft actuator to enable variable stiffness. The system's behavior has been modeled in both linear and nonlinear states, utilizing both the strain energy theory of hyperelastic materials and a novel hysteresis identification technique based on the Prandtl-Ishlinskii method. Moreover, the results have been validated with experiments. By adding a layer jamming actuator to the soft actuator, the newly structured robot can increase its stiffness up to nine times when the layer jamming is activated. If the layer jamming is deactivated, the robot behaves like a typical soft actuator. Moreover, as the test results indicate, the strain energy-based method shows a 6.3% deviation from the actual behavior in the linear range, while it was unable to accurately characterize the actuator's behavior in nonlinear states. In contrast, hysteresis modeling displays an 8.5% deviation from experimental data in both linear and nonlinear states. Overall, the combination of the layer jamming and soft bending actuator has resulted in a more versatile manipulator whose behavior could be modeled and anticipated with adequate accuracy considering both modeling techniques.