Performance evaluation of sdre-based spacecraft attitude estimation and control

This work investigates the performances of several State-Dependent Ric-cati Equation algorithms for the estimation and control of attitude and attitude rates of a rigid-body spacecraft. The estimators are designed to process line-of-sight and gyros measurements corrupted by white noises and potentially drifting biases. The case of gyroless estimation is also addressed. The vector measurements are linear-in-quaternion with state-dependent white noises. The controllers implement single-loop and dual-loop approaches. Under equivalent conditions, the single-loop controller outperforms the dual-loop controller wether by requiring less control energy or by having a quicker pointing transient. The dual-loop approach however requires less computations and shows a less aggressive control transient. Estimation performances are the limiting factor in the proposed partial information closed-loop attitude and rate controllers. With angular deviations in the vector observations of 0.5 deg, with or without gyros, and neglecting perturbations, pointing performance levels of typically 0.1 deg are demonstrated via numerical simulations.