Analysis of environment coupled noise terms in inertial sensor data using Allan Variance



Journal Title

Journal ISSN

Volume Title



Stability in inertial sensor data is commonly measured by using Allan Variance as a method of analysis. Allan Variance (AVAR) is the two-sample variance, originally developed to measure stability in oscillators. Plotting the AVAR curve for inertial sensor data that is mounted on a mechanically stable platform can aid in determining systematic errors and performance parameters of these sensors. This paper looks into inertial navigation and principles that govern the operation of these sensors. It will also explore various errors that affect the performance of these sensors, and how errors can be detected from the sensor data. It will then dive into the theory behind Allan Variance, and how this tool can help in analyzing and differentiating various noise terms that show up from inertial measurement data. An experiment is performed by using an Analog Devices tactical grade Inertial Measurement Unit (IMU) to observe the AVAR curves in different test setups of varying mechanical stability, and an analysis is done in terms of identifying various environment-coupled noise terms from the data.



Allan Variance (AVAR), Inertial Measurement Unit (IMU), Mechanical Stability, Navigation