Monocular vision planar motion decomposition measurement model-based dynamic calibration for IMU linear and angular sensitivity parameters

Inertial measurement units (IMUs) consist of multiple basic linear and angular motion measurement units and are widely used in fields such as pose estimation, motion control, virtual reality, vibration monitoring, etc. To ensure their measurement performance in engineering applications, it is necessary to regularly calibrate the linear and angular sensitivity parameters for each of their axes. However, the current methods struggle to balance the practical requirements for efficient, flexible, and accurate calibration, inevitably resulting in high time and economic costs. To address these issues, this study proposes a synchronous dynamic calibration method for all axes of IMUs by integrating the monocular vision with an orthogonal decomposition measurement model of planar motion, which can significantly improve efficiency and reduce costs, and also avoid the repeated installation errors. This method selects appropriate planar motions to synchronously provide excitations for each axis, and accurately reproduces these excitations by the planar motion decomposition measurement model and high-accuracy monocular vision method, achieving all linear and angular sensitivity parameter calibrations with low-cost equipment and simple steps. The experimental results compared with the traditional single-axis sequential calibration method confirmed that the calibration deviations of linear and angular sensitivities are 0.8% and 0.6% in the range of 0.01-5 Hz, respectively, and the overall calibration efficiency is improved by more than three times.