Kalman filter configurations for a low-cost loosely integrated inertial navigation system on an airship

A comparative study of various Kalman filter configurations applied to a low-cost inertial navigation system (INS) for an unmanned airship or blimp was made. The attitude of the airship was determined using micro-electromechanical sensors (MEMS) gyros with feedback from the earth's magnetic field and gravity vectors. MEMS accelerometer measurements were used as a gravity vector, as the dynamic accelerations on an airship are small compared with the static gravitational acceleration. The velocity and position estimates were updated from a loosely integrated GPS receiver. It was found that the best trade-off between accuracy and processing power could be achieved by having two smaller extended Kalman filters (EKF) running in sequence. The first EKF estimates the attitude of the airship, while the second EKF estimates the velocity and position of the airship. The solution was implemented on an onboard computer to provide real-time navigation estimates.