Multiple IMU sensor fusion using resolution refinement method to reduce quantization error

We propose a sensor fusion method of multiple inertial measurement units (IMU) with different resolutions to reduce quantization errors and improve the measurement accuracy of dead reckoning navigation. The resolution and dynamic range of the accelerometer within MEMS IMU have a trade-off relationship; the selection of a fixed dynamic range in a single IMU may increase the error of the inertial navigation system. We integrate multiple accelerometer signals with different resolutions into a weighted average that depends on the measurements. This reduces the quantization error by creating an optimal resolution tailored to the specific measurement range. To verify the proposed algorithm, we examined the effect of mathematically generated accelerometer values.