IEKF-based Self-Calibration Algorithm for Triaxial Accelerometer

Abstract

This paper proposed a self-calibration algorithm for triaxial accelerometer. By analyzing the measurement error factors, the parametric model of accelerometer output was built. According to the principle that the modulus value of gravity vector at a fixed point is constant, the nonlinear state space model of calibration parameters was derived. Further, the iterated extended kalman filter was proposed to avoid the problem of high space complexity of off-line calibration algorithm. Through numerical simulation the efficiency of the proposed iterated extended kalman filter algorithm was illustrated. Simulation results also demonstrated the superior performance of the proposed algorithm over the Least squares algorithm in the application of accelerometer calibration.