Normal mode attitude determination algorithm extended Kalman filtering approach

Abstract

An attitude determination system has been designed for Hughes future geosynchronous orbit (GEO) satellites which employ a continuously running inertia! rate sensor in conjunction with attitude sensors such as sun sensor, earth sensor, and/or star tracker. Real-time knowledge of spacecraft attitude with respect to an inertial reference frame is computed by numerically integrating gyro data through the relationship of kinematics equations. Attitude sensor data are processed to generate corrections to spacecraft attitude, gyro rate bias, and attitude sensor bias estimates. An extended Kalman filter based attitude determination algorithm is derived in this paper for a GEO satellite using static earth sensor assemblies (ESAs) and a two-axis fine sun sensor (FSS) as attitude sensors. A Matlab-based time domain simulation model is developed to evaluate the attitude determination performance. Simulation results show that precision attitude determination (less than ± 0.05 deg per axis) during normal mode operation is achieved using the selected attitude hardware and algorithms. 1.0 Introduction Attitude Determination Algorithm Overview The goal is to design an Attitude Determination and Control System (ADCS) which provides the spacecraft inertial attitude with accuracy better than ± 0.05 deg in roll, pitch, and yaw to meet the overall pointing requirement during the normal Earth pointing mode. The selected attitude hardware for precision attitude determination consists of (1) a three-axis gyro mounted on a radiator panel, (2) the two axis FSS mounted on the aft structure for sun viewing near sunset, (3) the static ESAs mounted at three places on the nadir structure to optimize Earth viewing angles, and (4) the Spacecraft Control Processor (SCP) which hosts the ADCS software and all sensor and actuator input/output (I/O) functions. Copyright @ 1997 by the American of Aeronautics and Astronautics, Inc. All rights reserved. This paper describes the design and performance evaluation of precision attitude determination algorithms using the aforementioned attitude sensors. A functional block diagram of the attitude determination algorithms is shown in Figure 1. The algorithms are partitioned into two major modules : attitude propagation and attitude estimator. Listed below are the nomenclature used throughout the paper: B co : a 3x1 vector representing three angular rates of a Bjody reference frame with respect to an Inertial reference frame, where the vector is expressed in Body reference frame. C : a 3x3 direction cosine matrix representing the B attitude of an Orbital reference frame with respect to a Body reference frame. Q : a quaternion presented by a 4x1 vector specifying B the rotation of an Qrbital reference frame with respect to a Bpdy reference frame. 1.1 Attitude Propagation. HIRU data is periodically sampled (every 32 msec.) into the SCP, compensated for known measurement errors such as input axis misalignments, converted into body reference frame, and corrected for gyro biases. The compensated gyro data is then used to numerically integrate a set of euler parameters (or quaternion) that specify the spacecraft attitude with respect to an orbital reference frame. 1.2 Attitude Estimator. Because of the initial attitude error used in quaternion integration and the accumulated effect of both gyro measurement errors and integration truncation and round off errors, attitude determined by processing only gyro data will contain errors that grow with time. The gyro measurement errors include gyro random walk, bias instability, resolution, and scale factor error. The function of the attitude estimator is to use Earth/Sun sensor data to estimate the spacecraft attitude determination errors. In addition, the attitude estimator estimates the gyro biases to be compensated in gyro