An LMI-based observer for attitude determination

In this paper we consider the problem of estimating the attitude of a vehicle using a Kalman filter based observer with increased robustness. This novel approach takes into account the nonlinearities and the uncertainties present in the system, by using a direct approach and the S-procedure, by the means of an estimator formulated as a linear matrix inequality problem. This is useful for dealing with navigation systems where the understanding and analysis of the nonlinearities are relevant for the final solutions. An experiment is carried with an integrated GPS/INS system mounted on a vehicle, in order to verify the performance of this estimator on a real case scenario.