摘要
针对倾转四旋翼无人机(TQUAV)在未知扰动下飞行的轨迹跟踪问题,提出了一种自适应循环神经网络(RNN)互补滑模控制(CSMC)方法。首先,将倾转四旋翼无人机动力学系统分为位置子系统和姿态子系统并分别设计互补滑模控制器。考虑到无人机在飞行过程中受到外部扰动的影响,由滑模控制方法得到的理想控制器包含未知的干扰项,因此使用循环神经网络来估计未知扰动。然后,为降低循环神经网络组合估计误差的影响并削弱控制器抖振,引入超螺旋滑模控制(STSMC)设计切换控制器。最后,通过仿真验证了所提方法的有效性。
An adaptive complementary sliding mode control(CSMC)based on recurrent neural network(RNN)is proposed to solve the trajectory tracking problem of tilt quadrotor unmanned aerial vehicle(TQRUAV)flying under unknown disturbances.Firstly,the dynamic system of the TQRUAV is divided into position subsystem and attitude subsystem,and complementary sliding mode controllers are designed sepa⁃rately.By considering the influence of external disturbances on the flight process of TQRUAV,the ideal con⁃troller obtained by sliding mode control contains unknown disturbance subjects.Therefore,the RNN is used to estimate the unknown disturbances.Then,in order to reduce the impact of combined estimation errors of RNN and weaken controller chattering,a super⁃twisting sliding mode control(STSMC)is introduced to de⁃sign switching controllers.Finally,the effectiveness of the proposed method is verified through simulation.
作者
李晨
熊晶晶
LI Chen;XIONG Jingjing(School of Mechanical and Electrical Engineering,China Jiliang University,Hangzhou 310018,China)
出处
《航天控制》
CSCD
2024年第1期30-36,共7页
Aerospace Control
基金
浙江省自然科学基金资助项目(LQ21F030016)。
关键词
倾转四旋翼无人机
互补滑模控制
自适应控制
循环神经网络
Tilting quadrotor UAV
Complementary sliding mode control
Adaptive control
Recurrent neural network
作者简介
李晨(1997-),男,硕士研究生,主要研究方向为多旋翼无人机智能控制方法;通信作者:熊晶晶(1987-),男,博士,讲师,主要从事非线性系统控制和时滞系统稳定性分析与同步控制等方面的教学与科研工作,E⁃mail:xjj286@cjlu.edu.cn。
