UAV의 자세 제어를 위한 개선된 슬라이딩 모드 제어

Abstract

This thesis introduces an improved sliding mode control technique for attitude control of UAVs with a quadrotor structure. The sliding mode control is a non-linear control technique that is robust against external disturbances and model uncertainty. To design a controller for UAVs, we define sliding variables and reaching laws and derive control inputs. To analyze whether the designed controller is stable in the presence of external disturbances, the stability is confirmed using the Lyapunov theory, and the control performance is verified through computer simulation.

Sliding mode control has two problems, such as chattering, which is a high-frequency oscillation, and it takes an infinite amount of time for the state to converge to the target. We discuss the improved reaching law to reduce chattering, and verify that chattering is reduced through simulation. To solve the second problem, we introduce terminal sliding mode control and fast terminal sliding mode control that guarantee a finite convergence time and verify their performance through simulation.

The improved reaching laws are effective in reducing chattering but have the disadvantage that robustness is not guaranteed. We introduce a super twisting algorithm that can reduce chattering while ensuring robustness. This control technique is one of the second-order sliding mode controls, and the control input is obtained by differentiating the sliding variable twice. The past algorithm guarantees only the stability of a model with one state variable, but in this work, the stability of a model with multiple state variables was confirmed. The designed controller has verified its control performance through simulation.

As UAVs perform various and complex missions, they can sometimes face various constraints. The mission can be performed in an environment where there is a limit on the size of the control input due to the physical limit of the actuator, or the angular rate limit due to ground shooting, transportation of important objects, etc. In a constrained environment, it is difficult to obtain optimal control performance with the existing control technique. Therefore, a control method considering constraints is required, and in this paper, a sliding mode-based control method is introduced. First, we describe model predictive sliding mode control for UAVs attitude control with a control input limit. This technique is a combination of model predictive control and sliding mode control that predicts the future state and reflects it in the current state. Next, we introduce constrained sliding mode control for the attitude control of UAVs considering the angular rate limit. This technique defines a new sliding variable with two equilibrium points and has a variable structure that switches according to the magnitude of the angular rate. Finally, Analyze the stability of the control technique and verify its performance through simulation.