Federated Nonlinear Predictive Filtering for the Gyroless Attitude Determination System with Multiple Sensors

In this paper, a federated nonlinear predictive filter is derived for the gyroless attitude determination system with multiple sensors (e.g., star sensor, sun sensors, GPS sensor). The features of this approach include a nonlinear predictive filtering algorithm to determine the attitude from the attitude dynamic model of the spacecraft without rate gyros and a federated filtering algorithm to process the measurement information from the multiple attitude sensors. The equivalence relation between nonlinear predictive filter and Kalman filter is demonstrated from algorithm structure and estimation criterion. The approach combines the good qualities of both the nonlinear predictive filter and federated filter. Simulation results using this algorithm indicate the filter accurately estimates the attitude of the spacecraft with the utilization of the star sensor and GPS sensor.