Iterative Learning Cascade Trajectory Tracking Control for Quadrotor-UAVs With Finite-Frequency Disturbances

作     者：Qian, Shiyu Xu, Jing Niu, Yugang Jiao, Ticao 

作者机构：East China Univ Sci & Technol Key Lab Smart Mfg Energy Chem Proc Minist Educ Shanghai 200237 Peoples R China Shandong Univ Technol Sch Elect & Elect Engn Zibo 255000 Peoples R China 

出 版 物：《IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY》 (IEEE Trans. Veh. Technol.)

年 卷 期：2025年第74卷第4期

页      面：5624-5636页

核心收录：

学科分类：0810[工学-信息与通信工程] 0808[工学-电气工程] 08[工学] 0823[工学-交通运输工程] 

基　　金：National Natural Science Foundation of China [62333006, 62173141] Natural Science Foundation of Shanghai [22ZR1417900] Taishan Scholar Project of Shandong Province [TSQN202306191] 

主　　题：Attitude control Robustness Quadrotors Uncertainty Trajectory tracking Tracking loops PD control Vectors Symmetric matrices Noise Sliding mode control cascade PD-PID control multi-loop control iterative learning control H-infinity control 

摘      要：Due to the under-actuated and strong coupling characteristics of the quadrotor system, the conventional trajectory tracking control methods always suffer from the low control precision, and the poor anti-disturbance capability. In this work, an iterative learning high-performance trajectory tracking control method is designed for quadrotors with finite-frequency uncertainties. The designed closed-loop flight control system includes the cascade connection of position and attitude loops. First, a cascade position loop, composed of the finite-frequency $H_\infty$ based PD-PID controllers, controls the translational dynamics of the quadrotor, and generates the desired attitudes for inner loops. Then, an iterative learning P-sliding mode controller is designed for inner loops to compensate for the repeatable tracking errors of the attitudes, and attenuate the sensor noises during the periodic flight maneuvers. By learning from previous iterations, the combination of SMC and PD-type learning terms yields a satisfactory inner-loop robust performance. The resulting algorithm requires little computational power and memory, and its convergence properties under measurement noise are shown. Experimental results have demonstrated the effectiveness of this approach in comparison to cascade PID-PID control and cascade PID-SMC methods.