Attitude Heading Estimation of Indoor Moving Object Based on Extended Kalman Filter Algorithm

Inertial navigation based on Micro-Electro-Mechanical System (MEMS) sensors has recently received much attention. Attitude heading estimation of an indoor moving object, as one of the important parts of inertial navigation is still worth studying. Attitude heading estimation using the acceleration and magnetic sensors fixed on the indoor moving object can easily be disturbed by surrounding environments or other factors. The gyroscope can also be used for computing the attitude heading. However, due to its own characteristics that the estimated errors gradually accumulates, the accuracy is not reliable for a long time. In this paper, quaternion-based Extended Kalman Filter algorithm is proposed to estimate the attitude heading of indoor moving object combining the acceleration sensor, magnetic sensor and gyroscope. For the method, the quaternion derived from gyroscope's readings is selected as the state vector of the filter, and the quaternion calculated from the measurements of accelerometer and magnetometer is used for the observations of the filter. The real-time estimated results are achieved by the state update of the designed filter. The experiment result shows that the proposed method can effectively improve the accuracy of the attitude heading estimation, compared with the complementary filtering.