SPACECRAFT ATTITUDE ESTIMATION USING THE REGULARIZED PARTICLE FILTER WITH ROUGHENING
William Silva1; Roberta Garcia2; Hélio Kuga1; Maria Zanardi3
1 Instituto Tecnológico de Aeronáutica (ITA); 2 Universidade de São Paulo (USP); 3 Universidade Federal do ABC (UFABC)
doi:10.20906/CPS/CON-2016-0706
Resumo
In this work, it will be described the attitude determination and the gyros drift estimation using the Regularized Particle Filter with Roughening for non-linear systems. The Particle Filter is a statistical, brute-force approach to estimation that often works well for problems that are difficult for the conventional Kalman Filter. The particle filter does not assume any statistical distribution of the phenomenon and therefore in real time applications the estimation accuracy and efficiency are significantly affected by numbers of particles chosen to represent the dynamics. The Regularized Particle Filter was used for preventing the sample impoverishment, which occurs when the region of state space in which the posteriori probability density function has significant values that does not overlap with the a-priori probability density function. This means that if all of our a-priori particles are distributed according to the a-priori probability density function, and compute the posteriori probability density function to resample the particles, only a few particles will become meaningful a-posteriori particles. The application uses the real measurement data for orbit and attitude of the CBERS-2 (China Brazil Earth Resources Satellite). The attitude dynamical model is described by nonlinear equations involving the Euler angles. The attitude sensors available are two DSS (Digital Sun Sensors), two IRES (Infra-Red Earth Sensor), and one triad of mechanical gyros. The two IRES give direct measurements of roll and pitch angles with a certain level of error. The two DSS are mounted on the satellite body such that they are nonlinear functions of roll, pitch, and yaw attitude angles. The gyros are aligned in the 3 satellite axes and furnish the angular measurements in the body frame reference system. The results in this work show that, by using the particle filter approach, can reach accuracies in attitude determination within the prescribed requirements, besides providing estimates of the gyro drifts which can be further u
Palavras-chave: Regularized Particle Filter; Roughening Technique; Unscented Kalman FIlter; nonlinear system; attitude determination