Navigation Observer Design Using Vector Measurements and a GPS Sensor

This article addresses the problem of designing a nonlinear navigation observer for quadrotor unmanned aerial vehicles (UAVs) using vector measurements and a GPS sensor. The observer has a cascaded structure and consists of a rotational state (angular velocity and attitude) observer and a translational state (position and linear velocity) observer. Based on the Lyapunov analysis, it is shown that the angular velocity observer is globally exponentially stable, the attitude observer is almost globally asymptotically stable, and the translational state observer is globally exponentially stable under a mild assumption on the attitude estimation error. The cascaded structure facilitates convergence analysis and observer implementation. The performance of the proposed observer was compared with that of an angular velocity observer reported in the literature in numerical simulations and validated through experiments.