Key points are not available for this paper at this time.
SLAM (Simultaneous Localization And Mapping) in unmanned aerial vehicles can be an advantageous proposition during dangerous missions where aggressive maneuvers are paramount. This paper proposes to achieve it for a quadcopter using a differential flatness-based linear quadratic regulator while utilizing sensor measurements of an inertial measurement unit and light detection and ranging considering sensors’ constraints, such as a limited sensing range and field of view. Additionally, a strategy to reduce the computational effort of Extended Kalman Filter-based SLAM (EKF-SLAM) is proposed. To validate the performance of the proposed approach, this work considers a quadcopter traversing an 8-shape trajectory for two scenarios of known and unknown landmarks. The estimation errors for the quadcopter states are comparable for both cases. The accuracy of the proposed method is evident from the Root-Mean-Square Errors (RMSE) of 0.04 m, 0.04 m/s, and 0.34 deg for the position, velocity, and attitude estimation of the quadcopter, respectively, including the RMSE of 0.03 m for the landmark position estimation. Lastly, the averaged computational time for each step of EKF-SLAM with respect to the number of landmarks can help to strategically choose the respective number of landmarks for each step to maximize the use of sensor data and improve performance.
Rauniyar et al. (Fri,) studied this question.