Laser-aided Navigation with Loop Closure Capabilities for Micro Aerial Vehicles in Indoor and Urban Environments

摘要

This paper presents a laser-aided navigation system for Micro Aerial Vehicles. It is based on a Kalman filter so that GNSS measurements can be incorporated if available. For GNSS-denied areas, the Kalman filter also processes relative pose measurements extracted from laser data. A novel approach for laser-aided Kalman filter navigation is presented which allows using multiple reference scans simultaneously. Furthermore, an addition to avoid growth of the heading angle error is described. Because this Kalman filter based system is a relative navigation system, its position error grows with time. To avoid such an error growth in GNSS denied environments, the Kalman filter is augmented with loop closure detection. A technique is proposed to represent such information in a pose graph and to calculate an improved navigation solution including an error covariance based on covariance intersection. The successful operation of the presented system is validated in several experiments including real flight data, large loops and an outdoor-indoor-outdoor transition.