Orientation Estimation Based on LKF Using Differential State Equation

Smart devices, such as mobile phones, smart watches, and other wearable devices, require accurate methodology to detect change in orientation. Determining the orientation of the smart device is useful for developing applications in the field of fitness, healthcare, sports, and gaming. There are existing methods based on sensor fusion to determine the orientation of a device, but these methods have drawbacks in terms of accuracy, stability, or response times. This paper proposes a new quaternion-based sensor fusion approach for estimating orientation based on accelerometer, magnetometer, and gyroscope sensors. The proposed method uses a low complexity linear Kalman filter based on the differential state equation and provides the substantial performance improvements by compensating for the sample-wise gyroscope drift in the time update system instead of the measurement update system. This new approach ensures that the uncertainty while using the Kalman filter is reduced and the orientation computed is both accurate and fast for use in smart device applications. Based on the improvements during orientation estimation, it is observed that the proposed solution detects minor orientation changes accurately with fast response characteristics and also completely remove any drift. The proposed method for the computation of orientation results in substantial improvements in response time and settling time compared with existing orientation estimation methods.