Design and Experimental Validation of Sensorless 4-Channel Bilateral Teleoperation for Low-Cost Manipulators

TL;DR

This paper presents a sensorless 4-channel bilateral teleoperation system for low-cost manipulators, enabling high-speed, stable, and contact-rich operations with force feedback, and enhances imitation learning success.

Contribution

It introduces a novel sensorless bilateral control framework with nonlinear dynamics compensation and a disturbance observer, reducing tuning complexity and improving performance in low-cost manipulators.

Findings

Stable high-speed teleoperation demonstrated in experiments.

Force feedback improves imitation learning success.

Reduced tuning complexity with a single cutoff frequency.

Abstract

Teleoperation of low-cost manipulators is attracting increasing attention as a practical means of collecting demonstration data for imitation learning. However, most existing systems rely on unilateral control without force feedback, which limits performance in fast or contact-rich operations under severe sensing and bandwidth constraints. This paper demonstrates that practical high-speed bilateral teleoperation with force feedback is achievable on force-sensorless, low-cost manipulators by employing a sensorless 4-channel bilateral control framework. The proposed method integrates nonlinear dynamics compensation with a disturbance-observer-based velocity and external force estimation scheme, enabling stable position-force interaction while avoiding the performance degradation caused by phase-lagged velocity estimation commonly used in low-cost systems. By interpreting the observer structure in the frequency domain, we clarify the intrinsic coupling between velocity and external force estimation bandwidths and show that the observer tuning freedom can be reduced to a single cutoff frequency, providing practical, hardware-oriented parameter tuning guidelines for low-cost implementations. Real-robot experiments demonstrate stable and accurate teleoperation in high-speed and contact-rich scenarios. Furthermore, as an application, we show that incorporating force information in demonstrations collected with the proposed system significantly improves the success rate of imitation learning across multiple manipulation tasks.