Abstract: The coarse sun sensor is an analog sun vector detector consisting of multiple solar cells. With its simple structure and low cost, the sensor is widely used in the platform of pico-satellites. The relationship between the photocurrent and sun light intensity is the fundamental characteristic of the coarse sun sensor. Generally, a cosine curve is used to fit that relationship, but it can cause a large measurement error. By analyzing experimental data, this article established an accurate mathematical model of the solar cell output characteristic, and developed a real-time on-orbit calibration algorithm. The algorithm has a multilevel coupling structure with one extended Kalman filter (EKF) and six Kalman fiters (KFs). It can estimate satellite attitude, angular velocity, and up to 30 different coarse sun sensor calibration parameters. Simulation shows that, compared with the commonly used EKF attitude estimator, the accuracy of the suggested algorithm is improved by about ten times, while the running time is only increased to approximately 1.5 times.

Key words: pico-satellite, coarse sun sensor, attitude estimation, sensor calibration, Kalman filtering