Multi-level variable dimension Extended Kalman Filter algorithm for UAV

To solve the problem of autonomous position and navigation for UAV(Unmanned aerial vehicle) in complex environment, this paper has proposed a Multi-level variable dimension extended Kalman filter fusion algorithm(ML-VDEKF). The motion and observation equations of the filter system have been established, and some low-cost sensors such as IMU(Inertial measurement unit), magnetometer, GPS and barometer have been combined to estimate state information during the flight. The actual experimental data shows that the proposed algorithm can provide accurate attitude, velocity and position. At the same time, the algorithm greatly reduces the computation when running in real-time operation of embedded devices, and can meet the flight requirement.