摘要
当无人机行进轨迹内存在明显转向行为时,若不能实现姿态角与响应曲线的有效耦合,则会使飞行器的抗扰能力下降,从而降低无人机飞行品质;为解决上述问题,设计基于机对抗网络的无人机串级线性自抗扰控制系统;按需连接串级线性跟踪微分器、自抗扰型无人机姿态控制器与行进位置控制器,完成无人机串级线性自抗扰控制硬件系统的设计;建立机器学习模型对抗网络,求解无人机串级线性动力学运动公式,联合相关运动数据,完成无人机串级线性动力学性能分析;定义串级线性位姿坐标,通过推导自抗扰运动节点矩阵的方式,计算具体的自抗扰性控制条件,实现对无人机串级线性位姿的自抗扰性控制,联合相关应用部件结构,完成系统设计;实验结果表明,所设计控制系统作用下,俯仰角、滚转角两类姿态角与标准响应曲线之间的耦合误差均不超过10%,即便在行进轨迹内存在明显转向行为的情况下,应用该系统也可以实现姿态角与响应曲线的有效耦合,能够有效保障无人机飞行器的抗扰能力。
With obvious steering behaviors in the trajectory of UAV,if the effective coupling between attitude angle and response curve is not solved,it makes the anti-disturbance ability of UAV decrease,thus reducing the flying quality of UAV.In order to solve above problems,a UAV-cascaded linear active disturbance rejection control(ADRC)system based on the machine countermeasure network is designed.The cascaded linear tracking differentiator,ADRC attitude controller and traveling position controller are required to complete the design of UAV-cascaded linear ADRC hardware system.The machine learning model countermeasure network is established to solve the UAV-cascaded linear dynamic motion formula and achieve its dynamic performance analysis by combining relevant motion data.The cascaded linear pose coordinates are defined,and the specific ADRC control conditions are calculated by deducing the ADRC motion node matrix,so as to realize the ADRC control of the UAV-cascaded linear pose,and complete the system design by combining the relevant components.Experimental results show that under the application of the designed control system,the coupling errors between the pitch angle,roll angle and the standard response curve does not exceed 10%,with a noticeable steering within the trajectory.Therefore,the system can ensure the effective coupling between attitude angle and response curve,effectively ensuring the anti-interference ability of UAV.
作者
张寒冰
ZHANG Hanbing(Institute of Artificial Intelligence,Zhejiang Dongfang Polytechnic,Wenzhou 325000,China)
出处
《计算机测量与控制》
2024年第9期170-176,共7页
Computer Measurement &Control
关键词
机器学习
无人机
串级线性自抗扰
跟踪微分器
姿态控制器
动力性能
machine learning
UAV
cascaded linear active disturbance rejection
tracking differentiator
attitude controller
dynamic performance
