Monocular SLAM Position Scale Estimation for Quadrotor Autonomous Navigation

Monocular Simultaneous Localization and Mapping (SLAM) provides a position estimation with an unknown scale factor. The scale factor depends on the initial conditions of the SLAM algorithm so that it is always different. This paper presents a scale factor estimator based on the Immersion and Invariance technique for quadrotor autonomous navigation. The estimator assumes that the quadrotor's linear velocity, attitude and the applied total thrust are measurable. The estimator performance is validated using numerical simulations and experiments.