基于反步滑模的四旋翼无人机飞行姿态自适应控制

Adaptive control of flight attitude for quadcopter unmanned aerial vehicles based on backstepping sliding mode

摘要:

四旋翼无人机应用频率不断增加，应用场景愈加复杂，对其飞行姿态控制性能提出了更高的要求。现有方法对四旋翼无人机进行位姿控制后，可能会导致无人机的飞行行为出现发散或振荡，无法达到期望的位置与姿态，因此基于反步滑模算法，设计四旋翼无人机飞行姿态自适应控制方法。根据四旋翼无人机结构，构建四旋翼无人机动力学模型，利用MEMS惯性传感器对四旋翼无人机当前位置与姿态进行测量，结合卡尔曼滤波算法对飞行姿态角进行解耦计算，依据四旋翼无人机飞行姿态控制需求，联合反步滑模算法设计控制器，以飞行姿态角——滚转角、俯仰角与偏航角解耦计算结果作为控制器的输入量，控制器输出结果即为飞行姿态控制参量，实现四旋翼无人机飞行姿态的有效控制。实验数据证明:应用本文提出方法对四旋翼无人机飞行位姿进行控制后，控制结果与期望位置和期望位姿基本相同，滚转角、俯仰角与偏航角最大偏差不超过4°，具有较高的控制精度。

Abstract:

The application frequency of quadcopter drones is increasing, and the application scenarios are becoming more complex, which puts higher requirements on their flight attitude control performance. The existing methods for attitude control of quadcopter drones may lead to divergent or oscillatory flight behavior, making it difficult to achieve the desired position and attitude. Therefore, based on the backstepping sliding mode algorithm, a self-adaptive control method for the flight attitude of quadcopter unmanned aerial vehicles is designed. Based on the structure of a quadcopter drone, a dynamic model of the quadcopter drone is constructed. MEMS inertial sensors are used to measure the current position and attitude of the quadcopter drone, and the Kalman filtering algorithm is combined to decouple the flight attitude angle. According to the flight attitude control requirements of the quadcopter drone, a controller is designed using the backstepping sliding mode algorithm, which measures the flight attitude angle roll angle The decoupling calculation results of pitch angle and yaw angle are used as inputs to the controller, and the output results of the controller are the flight attitude control parameters, achieving effective control of the flight attitude of quadcopter unmanned aerial vehicles. Experimental data shows that after applying the method proposed in this article to control the flight posture of quadcopter unmanned aerial vehicles, the control results are basically the same as the expected position and posture. The maximum deviation between roll angle, pitch angle, and yaw angle does not exceed 4 °, indicating high control accuracy.