Bending soft robots must be structured and predictable to be used in applications such as a grasping hand. We developed soft robot fingers with embedded bones to improve the performance of a puppetry robot with haptic feedback. The manufacturing process for bone-inspired soft robots is described, and two mathematical models are reported: one to predict the stiffness and natural frequency of the robot finger and the other for trajectory planning. Experiments using different prototypes were used to set model parameters. The first model, which had a fourth-order lumped mass-spring-damper configuration, was able to predict the natural frequency of the soft robot with a maximum error of 18%. The model and the experimental data demonstrated that bone-inspired soft robots have higher natural frequency, lower phase shift, better controllability, and higher stiffness compared with traditional fiber-reinforced bending soft robots. We also showed that the dynamic performance of a bending soft robot is independent of whether water or air is used for the media and independent of the media pressure. Results from the second model showed that the path of a bone-inspired soft robot is a function of the relative lengths of the bone segments, which means that the model can be used to direct the design of the robot to achieve the desired trajectory. This model was able to correctly predict the trajectory path of the robot.

中文翻译：

---

仿骨弯曲软体机器人


弯曲软机器人必须是结构化且可预测的，才能用于抓握手等应用。我们开发了嵌入骨骼的软机器人手指，以提高具有触觉反馈的木偶机器人的性能。描述了仿骨软机器人的制造过程，并报告了两个数学模型：一个用于预测机器人手指的刚度和固有频率，另一个用于轨迹规划。使用不同原型的实验用于设置模型参数。第一个模型采用四阶集总质量-弹簧-阻尼器配置，能够预测软体机器人的固有频率，最大误差为 18%。模型和实验数据表明，与传统纤维增强弯曲软体机器人相比，仿骨软体机器人具有更高的固有频率、更低的相移、更好的可控性和更高的刚度。我们还表明，弯曲软机器人的动态性能与介质是水还是空气以及介质压力无关。第二个模型的结果表明，仿骨软机器人的路径是骨段相对长度的函数，这意味着该模型可用于指导机器人的设计以实现所需的轨迹。该模型能够正确预测机器人的轨迹路径。