Stochastic Approach for Modeling a Soft Robotic Finger with Creep Behavior

内容摘要

Soft robots have high adaptability and safeness which are derived from their softness, and therefore it is paid attention to use them in human society. However, the controllability of soft robots is not enough to perform dexterous behaviors when considering soft robots as alternative laborers for humans. The model-based control is effective to achieve dexterous behaviors. When considering building a model which is suitable for control, there are problems based on their special properties such as the creep behavior or the variability of motion. In this paper, the lumped parameterized model with viscoelastic joints for a soft finger is established for the creep behavior. Parameters are expressed as distributions, which makes it possible to take into account the variability of motion. Furthermore, stochastic analyses are performed based on the parameters' distribution. They show high adaptivity compared with experimental results and also enable the investigation of the effects of parameters for robots' variability.