Unified Robot and Inertial Sensor Self-Calibration

Area of Science:

• Robotics
• Sensor Fusion
• Calibration Techniques

Background:

• Independent calibration of robots and inertial measurement units (IMUs) is costly, requiring specialized equipment like automated test rigs and high-accuracy tracking systems.
• Current methods necessitate separate calibration processes for robots and IMUs, increasing time and expense.
• Existing approaches often rely on expensive, external measurement systems for accurate pose determination.

Purpose of the Study:

• To present a novel, cost-effective, and streamlined approach for the simultaneous calibration of robots and IMUs.
• To eliminate the need for expensive, external tracking systems and dedicated test rigs.
• To enable calibration by simply attaching an IMU to a robot's end effector and executing continuous motion.

Main Methods:

• Utilizes continuous-time batch estimation for statistically optimal solutions.
• Formulates the calibration as a nonlinear least squares problem under Gaussian assumptions.
• Analyzes the structure of the Jacobian matrix associated with the estimation problem.

Main Results:

• Demonstrates a method for simultaneous robot and IMU calibration using only the IMU and robot motion.
• Achieves calibration without external measurement systems, reducing cost and complexity.
• Validated through numerical simulations and experimental testing, showing comparable or superior results to traditional methods.

Conclusions:

• The proposed method offers a quick, easy, and inexpensive alternative for calibrating robots and IMUs concurrently.
• Continuous-time batch estimation provides an efficient and accurate solution for simultaneous sensor and robot calibration.
• This approach significantly lowers the barrier to entry for accurate robot and IMU calibration in various applications.