The Tobit-Unscented-Kalman-Filter-Based Attitude Estimation Algorithm Using the Star Sensor and Inertial Gyro...

Area of Science:

• Spacecraft navigation and control
• Estimation theory
• Sensor fusion

Background:

• Environmental factors necessitate data censoring in spacecraft star sensors, degrading traditional attitude determination algorithms.
• Existing algorithms struggle with attitude estimation accuracy during star sensor outages.

Purpose of the Study:

• To develop a high-precision attitude estimation algorithm for spacecraft.
• To address the reduced attitude determination ability caused by star sensor data censoring.

Main Methods:

• Established nonlinear state equations for an integrated star sensor and gyroscope navigation system.
• Improved the unscented Kalman filter's measurement update process.
• Applied the Tobit model to describe gyroscope drift during star sensor failure, calculating latent measurement values and deriving measurement error covariance.

Main Results:

• Computer simulations verified the proposed algorithm.
• The Tobit unscented Kalman filter demonstrated a 90% improvement in accuracy compared to the standard unscented Kalman filter during a 15-minute star sensor failure.
• The filter effectively estimates errors caused by gyro drift.

Conclusions:

• The proposed Tobit unscented Kalman filter provides a feasible and effective method for high-precision spacecraft attitude estimation.
• The algorithm successfully compensates for gyro drift during star sensor failures.
• This approach offers theoretical support for practical engineering applications in spacecraft navigation.