A uniformly semiglobally exponentially stable nonlinear observer for GNSS- and camera-aided inertial navigation

Abstract

Abstract-In this paper a nonlinear observer design for estimation of position, velocity, acceleration, attitude and gyro bias of an Unmanned Aerial Vehicle (UAV) is proposed. The sensor suite consists of an Inertial Measurement Unit (IMU), a Global Navigation Satellite System (GNSS) receiver, a video camera and an altimeter. The camera and machine vision can track features from the environment and calculate the optical flow. These data, together with those from the other sensors, are fed to the observer, that is proven to be uniformly semiglobally exponentially stable (USGES). The performance of the observer is tested on simulated data by assuming that the camera system can provide the necessary information.

Nonlinear Observers for GNSS- and Camera-Aided Inertial Navigation of a Fixed-Wing UAV

“…This paper is built on the authors' previous works [25]- [27], [28], [16], and proposes exponentially stable non-linear observers for the estimation of position, velocity, specific force, attitude, and gyro bias of a fixed-wing UAV. Exponential stability is important for systems that are exposed to environmental disturbances, measurement noise, and uncertain initialization, since it guarantees strong convergence and robustness properties [29].…”

“…The idea behind this is to find a replacement for magnetometers: even though they are commonly used on payloads, they suffer from high and variable levels of disturbances and require an often complicated calibration procedure before use. To the best of the authors' knowledge, [25] was the first time that OF was used in a non-linear observer for navigation purposes.…”

“…The authors of the present paper have previously used OF vectors from a single camera to calculate the normalized body-fixed velocity of the UAV, which was fed into the nonlinear observer as a reference vector [25]; the observer was then successfully tested on experimental data in [26]. In [27], OF was used in combination with epipolar geometry and compared with the results of [26].…”

Dead reckoning of a fixed-wing UAV with inertial navigation aided by optical flow

“…For the OF computations to be useful in the estimators, a transformation to body-fixed velocity is necessary. The transformation is motivated by [23], and the pinhole camera model is used [24]. The camerafixed coordinate system is related to the bodyfixed coordinate system as illustrated in the downward-looking camera is aligned with the body z-axis, and the focal point of the camera is assumed to coincide with the origin of {B}.…”

Cascade attitude observer for the SLAM filtering problem

“…Grip et al [11] proved that with a minor modification, semi-global exponentially stability for time varying reference vectors could be achieved. An extension of this work was done by Fusini et al [12], in which using optical flow and GNSS velocity provided new reference vectors. An attitude and gyro bias estimation scheme is presented in Batista et al [13], where the dynamics are given as an LTV system and solved with a Kalman filter.…”