Data processing algorithm of MEMS inclinometer based on improved Sage-Husa adaptive Kalman filter

In the actual MEMS inclinometer's data processing, there are some problems. Such as model error exists in dynamically modeling; the measured signals may be include outliers in complex environment and prior knowledge of the noise statistical rule is insufficient. In order to solve these problems, an improved Sage-Husa adaptive Kalman filter is proposed. According to the model error, it adds a weighting function to the step variance matrix of the filter algorithm after judging the filter whether abnormal or not, which is used to inhibit divergent of the filter. And with outliers' problems, to achieve the purpose of restraining outliers, it keeps up new information original nature by using a fixed function weighted in the new information sequence of the filter algorithm equation. Finally, the experiment results show that this method can improve the robustness of the filter, inhibit outliers, and at the same time, make the variance of the output signal of MEMS inclinometer one order of magnitude smaller.