Binocular visual pendulum angle detection and anti-swing control of quadrotor suspension system

Abstract

Different from the pendulum angle detection by the optical capture system in the indoor environment, it is difficult to detect the pendulum angle of quadrotor suspension system in the outdoor environment. This paper presents a pendulum angle vision measurement scheme based on the model of the quadrotor suspension system, and applied to the quadrotor anti-swing control scheme with feedforward controller. The theoretical expression of the coordinate relationship between the pendulum angle and the spherical symbol is obtained by establishing the mathematical model of the quadrotor suspension system. The coordinates of spherical symbol fixed to the payload are detected by the visual algorithm. Simulations and experiments show that proposed method can accurately identify spherical symbol in sufficient light, and the maximum measurement error of $\alpha$ angle is −0.024 rad and $\beta$ angle is −0.068 rad in the Gazebo simulation. Meanwhile, the results in MATLAB simulation indicate the anti-swing control scheme with feedforward controller can effectively restrain the payload’s swing.