An Onsite Calibration Method for MEMS-IMU in Building Mapping Fields

Area of Science:

• Robotics and Sensor Technology
• Geomatics Engineering
• Building Information Modeling (BIM)

Background:

• Light Detection and Ranging (LiDAR) is crucial for acquiring data in Building Information Modeling (BIM).
• Micro electromechanical (MEMS) based Inertial Measurement Unit (IMU) sensors are vital for robot-based mapping, but prone to installation errors.
• Systematic errors in MEMS-IMUs, such as biases and axial deviations, significantly impact building mapping accuracy.

Purpose of the Study:

• To analyze systematic errors in MEMS-IMUs during robot-based building mapping.
• To develop an effective error calibration model for improving MEMS-IMU performance.
• To enhance the accuracy of LiDAR-based building information acquisition.

Main Methods:

• Analysis of systematic errors including biases, scale errors, and axial installation deviation in MEMS-IMUs.
• Development of a novel error calibration model incorporating a new sampling method for plane deviation correction.
• Application of the least-squares method for calibrating gravity acceleration and completing the calibration process.
• Integration of the calibrated model with the Gmapping algorithm for building mapping.

Main Results:

• The proposed error calibration model significantly improves the accuracy of MEMS-IMU data.
• Prediction accuracy for yaw is increased by 1-2 degrees.
• LiDAR-based building mapping results demonstrate enhanced accuracy compared to previous methodologies.

Conclusions:

• The developed error calibration model effectively addresses systematic errors in MEMS-IMUs.
• The improved accuracy in sensor data and mapping provides a practical foundation for advanced Building Information Modeling.
• This research contributes to more reliable and precise automated data acquisition for BIM.