DIRECT COMPENSATION AND KALMAN FILTER COMPENSATION OF GRAVITY VERTICAL DEVIATION BASED ON GRID DATABASE

摘要

The gravity model of WGS84 is commonly used in the calculation of gravitational acceleration known as normal gravity vector, but it cannot meet the requirements for high-precision inertial navigation systems because of the gravity anomaly. In this paper, the effects of gravity vertical deviation on navigation accuracy are analyzed and the principle of compensation is studied based on vertical deviation grid database. Then two practical compensation methods are presented. First, the direct compensation method converts the vertical deviation obtained from a look-up table into a gravity model component, which is compensated in the specific force equation. Second, the Kalman filter open-loop compensation method regards the vertical deviation as the accelerometer bias, getting the current attitude, velocity, position error based on state propagation, and uses open-loop correction. The two methods are implemented based on a 2.5 '× 2.5' vertical deviation database. The results show that the compensation effects are basically the same and can effectively compensate the vertical deviation of gravity, and improve the accuracy of inertial navigation system.