Torque-sensorless control of a high-ratio, backdrivable Wolfrom-gearbox for safe human-centered robotics

Area of Science:

• Robotics
• Control Systems
• Mechanical Engineering

Background:

• Safe physical human-robot interaction (pHRI) is crucial for collaborative robots.
• Compact, high-ratio actuators are desirable for lightweight and efficient designs.
• Detecting external disturbances in high-ratio actuators without torque sensors is a significant challenge.

Purpose of the Study:

• To develop and validate a torque-sensorless control framework for high-ratio actuators.
• To rigorously quantify the limitations of detecting external disturbances in such systems.
• To enable safe human-robot collaboration using high-ratio, backdrivable actuators.

Main Methods:

• Introduced a control framework combining a double-inertia analytical model with a higher-order sliding-mode disturbance observer (HOSM-DOB).
• Developed an analytical model to determine disturbance-detection bandwidth.
• Integrated HOSM-DOB into an impedance controller with active inertia shaping.
• Validated experimentally using a single-degree-of-freedom testbed and controlled collisions.

Main Results:

• Achieved torque estimation errors below 10% for soft contacts.
• Demonstrated HOSM-DOB outperforming classical disturbance observers by 35% in high-frequency impact detection.
• Confirmed compliance with ISO/TS 15066 safety limits up to 7000 rpm motor speed.
• Revealed a fundamental trade-off between gear ratio, motor inertia, and sensing performance.

Conclusions:

• High-ratio, backdrivable actuators, when coupled with observer-based control, can achieve high torque density and safe interaction.
• The proposed framework offers a viable alternative to direct-drive and harmonic-drive solutions in human-centered robotics.
• Classical inertia-matching principles do not optimize impact detectability in these systems.