Inertial Measurement Units: Modeling and Calibration
in: Bilgisayar Bilimleri ve MühendisliğiAlanında Uluslararası Araştırmalar – III (International Research in the Field of Computer Scienceand Engineering), ASLAY FULYA, Editor, Eğitim Yayınevi, Konya, pp.5-75, 2025
- Publication Type: Book Chapter / Chapter Research Book
- Publication Date: 2025
- Publisher: Eğitim Yayınevi
- City: Konya
- Page Numbers: pp.5-75
- Editors: ASLAY FULYA, Editor
- Open Archive Collection: AVESIS Open Access Collection
- Ankara Yıldırım Beyazıt University Affiliated: Yes
Abstract
Inertial Measurement Units (IMUs) consist of various sensors, including accelerometers, gyroscopes, and magnetometers. Depending on the design choices of the manufacturer, an IMU may integrate one or more of these sensors, often incorporating multiple units to provide redundancy and enhance reliability. These sensors generate high-frequency data, typically ranging from 100 Hz to 1000 Hz, making them significantly faster than alternative perception sensors such as cameras, LiDARs, and depth sensors commonly employed in mobile robotics. Due to their high update rates and acceptable error bounds, IMUs are the primary choice for short-term state estimation. In particular, within the Kalman filter framework—widely implemented across mobile robotic systems—the process update step is driven by measurements obtained from accelerometers and gyroscopes. Consequently, IMUs are indispensable in mobile robotics applications. However, these sensors inherently exhibit noise and biases, which can lead to unreliable outputs if not properly calibrated or employed alongside appropriate mathematical filtering techniques, such as the Kalman filter family. This work provides a comprehensive analysis of IMUs and their constituent sensors, examining their working principles, sources of noise, calibration models, limitations, and strengths.