IMU Motion Capture Method with Adaptive Tremor Attenuation in Teleoperation Robot System.

Area of Science:

• Robotics
• Human-Computer Interaction
• Biomechanics

Background:

• Teleoperation systems enable remote task execution in unstructured environments.
• Non-intuitive controls (keyboard, joystick) and physiological tremor hinder teleoperation performance.
• Need for intuitive interfaces and tremor reduction in robotic teleoperation.

Purpose of the Study:

• To develop an intuitive control interface for teleoperation robots using wearable sensors.
• To improve attitude estimation accuracy and reduce physiological tremor during teleoperation.
• To enhance the overall performance and usability of human-robot systems.

Main Methods:

• Established a human arm kinematics model using two gForcePro+ armbands with inertial measurement units (IMUs).
• Developed a regression model for angular transformation to account for joint misalignment.
• Implemented a variable gain extended Kalman filter (EKF) fusing surface electromyography (sEMG) signals to attenuate physiological tremor.

Main Results:

• Achieved good attitude estimation accuracy, with an average angular Root Mean Square Error (RMSE) of 4.837° ± 1.433° compared to VICON optical capture.
• Demonstrated improved robot tracking performance and tremor reduction when tested with the xMate3 Pro robot.
• Validated the effectiveness of the developed control interface and tremor-filtering method.

Conclusions:

• The gForcePro+ armband-based control interface offers an intuitive and accurate method for teleoperation.
• The angular transformation model and EKF-based tremor reduction effectively enhance human-robot interaction.
• This approach shows significant potential for improving teleoperation in various applications requiring precise remote control.