Intelligent Real-Time MEMS Sensor Fusion and Calibration

What is it about?

This paper discusses an innovative adaptive heterogeneous fusion algorithm based on the estimation of the mean square error of all variables used in real-time processing. The algorithm is designed for a fusion between derivative and absolute sensors and is explained by the fusion of the three-axial gyroscope, three-axial accelerometer, and three-axial magnetometer into attitude and heading estimation. Our algorithm has a similar error performance in the steady-state but a much faster dynamic response compared with the fixed-gain fusion algorithm. In comparison with the extended Kalman filter, the proposed algorithm converges faster and takes less computational time. On the other hand, the Kalman filter has a smaller mean square output error in a steady state but becomes unstable if the estimated state changes too rapidly. In addition, the noisy fusion deviation can be used in the process of calibration. This paper proposes and explains a real-time calibration method based on machine learning working in the online mode during run time. This allows compensation of sensor thermal drift right in the sensor's working environment without the need for recalibration in the laboratory.

Featured Image

Photo by Luigi Frunzio on Unsplash

Photo by Luigi Frunzio on Unsplash

Why is it important?

Although the fusion method proposed by this paper is derived for this special case, the authors believe it can be used in many other applications as an alternative to the extended Kalman filter. The proposed fusion method is especially suitable if the sensor readings are recursively processed by non-linear functions. The proposed fusion and calibration methods can be adjusted to other sensor fusion scenarios (e.g. combination of the GNSS system – absolute velocity and a position sensor and accelerometer – a differential velocity sensor; an impulse volume sensor and a flow sensor and many others).