Enhancing the Force Transparency of TDPA-PP by Integrating Measured Force Feedback

Abstract

Teleoperation enables human intervention in remote robotic systems when autonomous control is unavailable or fails in unstructured or critical situations. Effective bilateral teleoperation requires high transparency in both position and force feedback. However, communication delays and artifacts, such as packet loss, can destabilize the system. Position-based time-delayed bilateral teleoperation architectures excel at precise position tracking but often suffer from reduced force transparency due to high damping during free motion. Motivated by novel developments in position-position architectures, the aim of this thesis is twofold: to improve a classical TDPA-PP architecture and compare with regarding force transparency, tracking accuracy, and stability.

TDPA-ER-PP ensures passivity of the two-port coupling, providing a robust bidirectional teleoperation framework. This thesis extends by integrating an additional force-feedback channel and by improving virtual-energy llocation, resulting in improved force transparency. Furthermore, a transfer from the current 1-Degree-of-Freedom case to the translation space of torque-controlled robotic arms is provided. The proposed modifications are experimentally validated on 1-Degree-of-Freedom devices, which are highly susceptible to instability, and further confirmed on a dual-arm 7-Degree-of-Freedom robotic setup.

In parallel to the well-established passivity-based methods, the Dual-Proxy model, proposes a novel approach that avoids the high damping of position-based architectures through task-space separation. However, the provided stability argument lacks a formal mathematical proof, and extensive filtering complicates objective evaluation. This motivates the independent analysis of the stability and transparency properties of the Dual-Proxy model presented in this thesis, via implementation on the same physical hardware used for the proposed modifications to TDPA-ER-PP.

Experimental results demonstrate that the methods proposed in this thesis outperform both TDPA-ER-PP and the Dual-Proxy model in force transparency while guaranteeing mathematical stability.