Integration of both actuation and proprioception into the robot body would provide actuation and sensing in a single integrated system. Within this work, a manufacturing approach for such actuators is investigated that relies on 3D printing for fabricating soft-graded porous actuators with piezoresistive sensing and identified models for strain estimation. By 3D printing, a graded porous structure consisting of a conductive thermoplastic elastomer both mechanical programming for actuation and piezoresistive sensing were realized. Whereas identified Wiener-Hammerstein (WH) models estimate the strain by compensating the nonlinear hysteresis of the sensorized actuator. Three actuator types were investigated, namely: a bending actuator, a contractor, and a three DoF bending segment (3DoF). The porosity of the contractors was shown to enable the tailoring of both the stroke and resistance change. Furthermore, the WH models could provide strain estimation with on average high fits (83%) and low RMS errors (6%) for all three actuators, which outperformed linear models significantly (76.2/9.4% fit/RMS error). These results indicate that an integrated manufacturing approach with both 3D printed graded porous structures and system identification can realize sensorized actuators that can be tailored through porosity for both actuation and sensing behavior but also compensate for the nonlinear hysteresis.

翻译：将驱动与本体感知集成于机器人本体中，可实现驱动与传感的一体化系统。本研究探索了一种通过3D打印制造具有压阻传感功能的梯度孔隙软体驱动器的工艺方法，并建立了用于应变估计的辨识模型。通过3D打印技术，利用导电热塑性弹性体构建梯度多孔结构，同时实现了驱动功能的机械编程与压阻传感。采用维纳-哈默斯坦（WH）模型通过补偿传感驱动器的非线性迟滞效应来估计应变。本研究开发了三种驱动类型：弯曲驱动器、收缩驱动器及三自由度弯曲段（3DoF）。结果表明，收缩驱动器的孔隙率可独立调控行程与电阻变化。此外，WH模型对三类驱动器的应变估计平均拟合度高达83%，均方根误差低至6%，性能显著优于线性模型（拟合度76.2%，均方根误差9.4%）。这些结果表明，结合3D打印梯度多孔结构与系统辨识的集成制造方法，既能通过孔隙率定制驱动与传感行为，又能补偿非线性迟滞效应，从而制备出传感化驱动器。