Gravity Vector Estimation from Integrated GPS/Strapdown IMU Data

Abstract

The Air Force is planning to launch a helium-filled balloon possessing a payload of an inertial ring laser gyro strapdown system and a 3-antenna, minimum 12-channel, dual-frequency GPS receiver. A ground-fixed, single-antenna GPS receiver will be centrally placed beneath the balloon’s trajectory, allowing for differential mode GPS tracking. The inertial measurement unit’s (IMU’s) accelerometers will measure the total kinematic accelerations in the body frame. These will be transformed to local (n,e,d) navigation frame components using the integrated attitude of the balloon obtained from the IMU’s gyroscopes. GPS-determined accelerations will reflect total inertial accelerations, which can easily be expressed in the navigation frame. Differencing the two sets will thus yield the desired gravitational accelerations. An elaborate covariance study was conducted involving a 36-state, 43-noise-process open-loop Kalman filter. Using conservative initial IMU and GPS error source values, the covariance study suggests that estimation of all three components of balloon-borne gravity vectors to an accuracy of 5 mGa1 (1 mGa1 = 10~5 m/s2) is feasible using “within reach” electronic and engineering capabilities. Leveling uncertainties are the main contributors to the overall gravity vector estimation error budget.