Estimation of bounded unknown input and its application to attitude estimation for accelerated rigid body using GPS/INS measurements

In this paper, we propose an observer based on proportional-integral type sliding mode to estimate unknown bounded time varying input of a nonlinear dynamical system. Exponential convergence of the observer is demonstrated using Lyapunov theory. Then this observer is used in the attitude estimation algorithms of accelerated rigid-bodies. The proposed observer is able to estimate the unknown linear acceleration using both linear velocity and specific acceleration obtained via GPS/INS measurements. Since this estimate is exponentially convergent, it can be used in most of attitude estimation algorithms. Simulation results are given to show the performance of this method and the importance of the estimate of linear acceleration to improve the estimation of the attitude of a rigid-body.