HYBRID INTRINSIC–EXTRINSIC CALIBRATION OF MULTIPLE IMUS AND JOINT ENCODERS FOR ACCURATE END-EFFECTOR TRAJECTORY ESTIMATION OF AERIAL MANIPULATORS

Ilyar Asl Sabbaghian Hokmabadi, Mahdis Bisheban

Keywords: Intrinsic calibration, Extrinsic calibration, Inertial Measurement Units (IMUs), Sensor fusion, Aerial manipulators, Robotic arm calibration, Optimization-based calibration

Abstract: Accurate pose and trajectory estimation of aerial robotic manipulators is an important area of research. The ubiquitous joint encoders commonly used in robotic arms suffer from limitations such as an inability to capture vibrations along all three axes of a link, failure to account for link deformation, and poor performance in detecting very fast motions. To address these issues, recent studies have equipped robotic arms with inertial measurement units (IMUs). However, achieving accurate IMU–encoder sensor fusion requires proper calibration. This study focuses on estimating intrinsic calibration parameters (e.g., IMU bias and scale), extrinsic calibration parameters (IMU-to-joint frame transformations), and robot model parameters (joint rotation axes and translations). The proposed system-wide calibration is performed without external equipment, relying instead on the gravity vector, static phases, and approximate forward kinematics of the robotic arm. Experiments conducted on a 7-DOF robotic arm equipped with low-cost IMUs demonstrate that simultaneous intrinsic–extrinsic calibration significantly enhances the accuracy of encoder–IMU fusion.

Submission Number: 26