A barometer-IMU fusion method for vertical velocity and height estimation

We propose a new fusing method to estimate the vertical velocity and the height of barometer-IMU systems. The method removed initialization processes of the system by calibrating accelerometer errors on-line. An EKF tracks the bias and scale error of each accelerometer, while estimating vertical velocity and height simultaneously. The method is tested on simulation and on an actual low-cost barometer-IMU sensor platform. The errors are successfully modeled before 30 to 60 seconds of system startup. Proposed method may be extensively used for low-cost, chip-scale barometer/IMU integrated devices, since on-line calibration not only removes initialization processes but also enables tracking of time/temperature dependent varying IMU errors from which MEMS based IMUs suffer.