Three-Finger Borescope-Mounted IPMC Gripper.

Ionic polymer-metal composites (IPMCs) are prominent soft actuators for millimeter-scale robotics due to their large bending deformation and high force-to-weight ratio. This work introduces an integrated soft robotic system featuring a three-channel IPMC gripper attached to a borescope for micromanipulation. A laser ablation technique partitions a single IPMC into three independently actuated channels, simplifying wiring complexity and enabling stable grasping of objects with complex geometries. While the hardware provides effective manipulation, teleoperation in microscale environments presents a significant challenge, as operators struggle to reliably determine contact and grasp stability from 2D visual feedback alone. To address this, we introduce a vision-based operator assistance system. Using a YOLOv8 segmentation model, our system processes the borescope's video feed in real-time to identify the gripper fingers and target objects, providing clear visual cues for "touched" and "grabbed" states. This human-in-the-loop feedback enhances operator precision and consistency and, for the first time, enables a quantitative evaluation of the gripper's performance. Experimental results demonstrate reliable grasping of diverse objects (0.3-6 mm) and validate the system's ability to provide objective success metrics. This work contributes an integrated and intelligent micromanipulation system that augments human capabilities and establishes a framework for future autonomous control.