A new method of model compensation in Kalman filter Application to launch vehicles
Abstract
The Kalman filter is commonly used to process tracking data in real time during satellite launching. Simple dynamic models are used due to stringent timing constraints. Due to mismatch in dynamics between real world and filter model, the filter diverges. This paper describes a new and simple method of estimating state noise covariance corresponding to the unmodeled part of the dynamics and using it to compensate the filter model to avoid divergence. This is based on processing perfect measurements and by proper transformation of the residuals. This method is applied for a typical satellite launch vehicle having maneuvers both in pitch and yaw planes.
- Publication:
-
14th Space Technology and Science Symposium
- Pub Date:
- 1984
- Bibcode:
- 1984spte.symp.1071P
- Keywords:
-
- Kalman Filters;
- Launch Vehicles;
- Spacecraft Launching;
- Covariance;
- Noise;
- Pitch (Inclination);
- Yaw;
- Launch Vehicles and Space Vehicles