Our proprioceptive studies serve as high-level research for improving dexterous telemanipulation. These studies focus on the interplay between accurate human hand tracking and haptic feedback rendered by a remote robotic system, and how this combination influences the quality of teleoperation.
Conducted with human participants, these experiments aim to understand individual preferences and behaviors when interacting with the telemanipulation framework. This is crucial, as different users may perceive and respond to haptic cues in unique ways.
Hand tracking accuracy and the influence on telemanipulation
The research has two primary goals:
Optimize hand tracking accuracy through calibration methods that estimate the relative position between the human hand and the robot’s hand, improving real-time motion alignment.
Evaluate the impact of tracking precision on robotic hand posture estimation and overall teleoperation performance.
Initial results show that recalibrating the system to better estimate the virtual link between human and robotic joints significantly improves tracking accuracy. Current studies are exploring how varying levels of tracking precision affect telemanipulation performance—especially when users view either their own hand or a standardized robotic hand in virtual reality.
The long-term goal is to uncover consistent human preferences and patterns that can guide the design of more intuitive, efficient, and reliable human-robot interfaces.
Exploring Haptic Rendering for Bilateral Telemanipulation
Maestro enables continuous 3D force rendering on the fingertips, providing high-precision haptic feedback during telemanipulation. However, the optimal haptic interaction pattern that enhances user performance and system reliability is still unknown.
While high-fidelity force feedback from the robot hand could increase the sense of immersion, it may also introduce unintended disturbances that reduce task efficiency or robustness. Additionally, the impact of task type and individual user preferences on haptic feedback quality over time remains an open question.