首页|期刊导航|控制理论与应用|数据优化PID参数整定的无人直升机自适应容错控制

数据优化PID参数整定的无人直升机自适应容错控制OA

Adaptive fault-tolerant control of an unmanned helicopter via data-optimized PID tuning

中文摘要英文摘要

针对具有执行器故障的无人机的位置控制与姿态控制问题,本文设计了一种数据优化PID的自适应容错控制器,通过引入直升机动力学模型保证系统的稳定性,并结合神经网络利用实时数据在线修正执行器故障的影响,构建模型与数据协同作用机制.首先,结合无人直升机的系统动力学模型,搭建高保真的无人直升机故障注入仿真平台;其次,基于无人直升机飞行过程中实时采集的状态数据与跟踪误差信息,并通过神经网络进行在线学习,实现基于数据的PID控制器参数在线优化与调节,在不依赖精确故障类型和故障程度等先验信息的情况下,有效补偿由执行器故障和模型不确定性引起的控制性能下降,提高系统对执行器故障的鲁棒性;最后,通过构建不同程度的执行器故障场景对本文设计的控制器的鲁棒性与稳定性进行验证.实验结果表明,本文提出的数据优化PID参数的控制器可根据无人直升机当前飞行状态和故障情况自适应调节增益,当出现执行器故障时比传统的PID控制器具备更优的自适应能力和控制性能,如三轴姿态的平均绝对误差显著降低约60%∼90%,部分故障情况下误差进一步收敛至10−4以下,具备更强的容错控制性能和鲁棒性.

This paper proposes a data-optimized PID adaptive fault-tolerant controller for the position and attitude con-trol of an unmanned helicopter with actuator faults.By combining the helicopter dynamic model with a neural network,the proposed method establishes a model-data cooperative control mechanism.The helicopter dynamic model is used to ensure system stability.The neural network is used to process real-time flight data and optimize PID parameters online.Firstly,a high-fidelity fault-injection simulation platform is built based on the system dynamics of the unmanned helicopter.Then,real-time flight states and tracking errors are collected during flight.These data are used for online neural-network learning.The neural network optimizes and adjusts the PID parameters adaptively.This process does not require prior knowledge of the exact fault type or fault severity.This enables the controller to compensate for performance degradation caused by actuator faults and model uncertainties so that improves system robustness.Finally,actuator fault scenarios with different fault severities are constructed to verify the robustness and stability of the proposed controller.Simulation results show the proposed controller which adaptively adjust its gains according to the current flight state and fault condition.Compared with conventional PID control,the mean absolute errors of the three-axis attitude are reduced by approximately 60%to 90%,and the errors converge to the order of 10−4,demonstrating stronger fault-tolerant performance and robustness.

郭黎;陈娟娟;李润泽;吴官正;沈成香

安徽工程大学 电气工程学院,安徽 芜湖 241000||安徽工程大学 高端装备先进感知与智能控制教育部重点实验室,安徽 芜湖 241000安徽工程大学 电气工程学院,安徽 芜湖 241000安徽工程大学 电气工程学院,安徽 芜湖 241000||安徽工程大学 高端装备先进感知与智能控制教育部重点实验室,安徽 芜湖 241000安徽工程大学 电气工程学院,安徽 芜湖 241000安徽工程大学 电气工程学院,安徽 芜湖 241000

无人直升机容错控制参数整定自适应控制执行器故障

unmanned helicopterfault toleranceparameter tuningadaptive control systemsactuator faults

《控制理论与应用》 2026 (9)

1929-1938,10

国家自然科学基金项目(U25A20453,62020106003),"直升机动力学"全国重点实验室开放基金项目(2024-ZSJ-LB-02-04,2024-CXPT-GF-JJ-093-08),江苏省自然科学基金项目(BZ2024037),安徽省科研编制重大项目(2023AH040121)资助. Supported by the National Natural Science Foundation of China(U25A20453,62020106003),the Open Fund of the National Key Laboratory of Helicopter Aeromechanics(2024-ZSJ-LB-02-04,2024-CXPT-GF-JJ-093-08),the Natural Science Foundation of Jiangsu Province of China(BZ2024037)and the Natural Science Research Project of Anhui Province Universities(2023AH040121).

10.7641/CTA.2026.60015

评论