Contingency Designs for Attitude Determination of Trmm

In this paper, several attitude estimation designs are developed for the Tropical Rainfall Measurement Mission (TRMM) spacecraft. A contingency attitude determination mode is required in the event of a primary sensor failure. The final design utilizes a full sixth-order Kalman filter. However, due to initial software concerns, the need to investigate simpler designs was required. The algorithms presented in this paper can be utilized in place of a full Kalman filter, and require less computational burden. These algorithms are based on filtered deterministic approaches and simplified Kalman filter approaches. Comparative performances of all designs are shown by simulating the TRMM spacecraft in mission mode. Comparisons of the simulation results indicate that comparable accuracy with respect to a full Kalman filter design is possible. Introduction The TRMM spacecraft is due to be launched in 1997 with a nominal mission lifetime of 42 months. The main objectives of this mission include: (1) to obtain multi-year measurements of tropical and subtropical rainfall, (2) to understand how interactions between the sea, air, and land masses produce changes in global rainfall and climate, and (3) to help improve the modeling of tropical rainfall processes and their influence on global circulation. The spacecraft is three-axis stabilized in a near circular (350 km) orbit with an inclination of 35°. The nominal Earth-pointing mission mode requires a rotation once per orbit about the spacecraft’s yaxis. The attitude determination hardware consists of an Earth Sensor Assembly (ESA), Digital Sun Sensors (DSS), Coarse Sun Sensors (CSS), a Three-Axis Magnetometer (TAM), and gyroscopic rate sensors. The attitude control hardware includes three Magnetic Torquer Bars (MTB) which are used to provide magnetic momentum unloading capability, and a Reaction Wheel Assembly (RWA) which consists of four wheels in a pyramidal arrangement to maximize momentum storage capability along a preferred axis. Primary attitude determination is accomplished using the ESA and gyroscopes. The DSS is also used twice each orbit in order to update the yaw position estimate during mission pointing. The allotted attitude knowledge accuracy is 0.18° per axis. Simulation studies indicate that the primary attitude determination system meets the knowledge requirements [1]. In the event of an ESA failure, a contingency mode is used to allow for the continuation of the scientific mission. Attitude determination for the contingency mode is accomplished using the DSS, the TAM, and gyroscopes. The allotted attitude knowledge accuracy for the contingency mode is 0.7° per axis. The algorithm chosen for the final contingency design incorporates a sixth-order Kalman filter [2]. This filter estimates both attitude error angles and gyro drift trajectories. However, due to initial concerns in software coding size and computations, the development of simpler and less softwareintensive algorithms is required. A number of algorithms is presented in this paper, including: an