A Performance Analysis Of Pose Estimation Based On Two-View Tracking And Multi-State Constraint Kalman Filter Fusion

This paper presents a performance analysis of two-view tracking and Multi-State Constraint Kalman Filter (MSCKF) fusion for a pose estimation. The system and measurement model of both two-view tracking and MSCKF are derived based on the fusion condition. The simulation result of the fused algorithm using the Drone Racing dataset, collected from an aggressive flight of micro aerial vehicle (MAV), shows the performance improvement of both attitude and position estimation compared to the performance of MSCKF.