Page 70 - IJAMD-1-1

International Journal of AI
for Material and Design Machine learning for gripper state prediction

research offers an in-depth examination of the gripper’s a cable-driven approach involving a motor and strings for
physical behavior under different actuation scenarios, movement activation. Notably, the cable was strategically
systematically overcoming the challenges of modeling positioned off-center from the joint to generate a bending
soft materials and complex joint interactions. Besides, moment when tensioned.
this work illustrates the potential of machine learning in The finger’s architecture incorporated a variety of
bridging the gap between theoretical models and practical materials, such as TPU and c-PLA, each fulfilling distinct
applications, thus pushing the boundaries of what can be roles. Predominantly made of TPU, the finger was designed
achievable in soft robotics. to interact gently with objects, reducing the risk of damage
2. Methods during gripping. The c-PLA material serves a dual purpose:
it functions as the structural backbone, which supports
2.1. Design of the gripper with stiffness-tunable significant loads and facilitates variable joint stiffness when
joints subjected to heat.
In this work, we investigated a gripper adopted from our A key feature of this gripper design is its capacity
previous work, which focused on a cable-driven, stiffness- for varying joint stiffness, offering enhanced actuation
13
tunable soft gripper made of c-PLA and TPU. The flexibility. This flexibility is achieved through multiple
proposed gripper design features an assembly of several joints in each finger, which can be individually controlled
fingers connected to a central mount, as illustrated in for stiffness, as illustrated in Figure 1C. The ability to
Figure 1A and B. To facilitate swift actuation, we employed modify joint stiffness exploits the special phase change

A C

B

D

E

Figure 1. Overview of the cable-driven stiffness-tunable gripper design and functionality. (A and B) The proposed design of the gripper’s finger highlights
the pulling direction of the nylon string and the integration of the conductive PLA backbone into the soft TPU finger. (C) The process of stiffness
modulation in the joints through electrical heating demonstrates the softened state of PLA when heated and its re-stiffening upon cooling, with an inset
thermogram showing temperature distribution during heating. (D) The fabrication process of the gripper’s finger depicts the sequence of TPU and c-PLA
extrusion. (E) Illustrations of flexible actuation achievable by the gripper, showing various configurations from no actuation to the actuation of both joints
(Copyright © 2023 John Wiley and Sons. Reprinted with permission from Goh et al. ).
13
Abbreviations: cPLA: Conductive polylactic acid; PLA: Polylactic acid; TPU: Thermoplastic polyurethane.

Volume 1 Issue 1 (2024) 64 https://doi.org/10.36922/ijamd.2328