Adaptive backstepping control for autonomous shipboard landing of a quadrotor with input saturation

Abstract

This paper presents a novel autonomous shipboard landing control algorithm of a quadrotor subject to external disturbances and input saturation. A coupled six-degrees-of-freedom (6-DOF) nonlinear relative motion model is derived to facilitate the controller design. To overcome the under-actuated property of the quadrotor and avoid collision, the shipboard landing process is cooperatively completed by a relative position controller (RPC) and a relative attitude-altitude controller (RAC). The RPC is proposed for use when the quadrotor is far away from the ship, aiming to drive the quadrotor to the vicinity of the ship. The RAC is used for the situation when the quadrotor is close enough to the ship to land the quadrotor on the ship steadily. An adaptive backstepping technique is proposed for both RPC and RAC, where unknown bounds of the lump disturbances are estimated online by the adaptive laws. To satisfy the input constraint requirement, a smooth model is employed to describe the control input saturation. Stability analysis proves that all states of the closed-loop systems are uniformly ultimately bounded. A numerical example verifies the performance of the control approach.

Robust neural network‐based backstepping landing control of quadrotor on moving platform with stochastic noise

“…Form Assumption 3, one concludes . Commonly, one can assume 11–13 However, this may degrade the convergence speed of the states to their desired values, especially when the quadrotor is far away from the ship. To alleviate this issue, it is assumed that the limitations of the control inputs are unknown, as .…”

“…However, the above‐mentioned approaches 5–10 only use the UAV dynamics in the controller design and the motion of the moving target was considered as disturbance inputs. In Reference 11, a coupled relative motion and attitude model was suggested and a robust command filter backstepping was then developed to assure the safe landing procedure. In Reference 12, an adaptive sliding mode control was suggested for the coupled 6‐degree‐of‐freedom relative system.…”

“…The approach of Reference 12 was then improved in Reference 13 by considering a terminal sliding mode control and fixed‐time stability. However, in References 11–13, the effect of stochastic noises on the system performance is not investigated.…”

“…Compared with state‐of‐the‐art methods the proposed approach has the following advantages, (I) in contrast with References 5–10, the proposed approach uses the ship dynamics explicitly. (II) The proposed approach is robust against stochastic noises compared to the approaches 11–13 . (III) Since an artificial redundant is used, the saturation limits are assumed to be unknown, however, in References 11–13, the limits of the control inputs are known.…”

“…(II) The proposed approach is robust against stochastic noises compared to the approaches 11–13 . (III) Since an artificial redundant is used, the saturation limits are assumed to be unknown, however, in References 11–13, the limits of the control inputs are known. The proposed controller is applied to the nonlinear quadrotor landing on a ship deck.…”