Quaternion Slerp for Attitude Estimation of Dual Spin-Stabilized Satellite

Satellite in orbit has a lot of challenges. Two of them are attitude stability and attitude determination. One method to stabilize a satellite is by spinning it, either the whole body (single-spin) or part of its body (dual-spin). However, this method has a side effect called nutation. In attitude determination, star-sensor has the best accuracy but is vulnerable to stray light that can make it blind. When the star-sensor is blind, the attitude needs to be estimated. The estimation becomes more complicated when nutation exists. The objective of this paper is to implement some methods to estimate this kind of satellite's attitude. First, four well-known methods, LinEuler, Lerp, Slerp, Squad, are implemented to LAPAN-A3's attitude data. This research found that three methods are failed to give a good estimation. Meanwhile, Slerp gives a considerable result with accuracy around 0.222° until 0.462° in the nutation direction (yaw and roll). Furthermore, two methods are developed to improve the estimation accuracy. Mean of multiple Slerp (mSlerp) is the first method. Using this method, the accuracy slightly improved, 0.203° until 0.361° in the nutation direction with a more consistent result compared to Slerp. However, neither Slerp nor mSlerp was able to estimate the nutation. Hence, the next method, Nutation Prediction Slerp, NPSlerp, is developed. This method can achieve far better accuracy until 0.007° by estimating the nutation. The drawback of NPSlerp is the accuracy increase as the length of interpolation increases.