目前机电作动器由于具有干净、维护方便等优点，越来越受到航空业的青睐。航空机电作动器的特点是控制精度、稳定性和响应速度要求高，针对以上特点，提出了一种基于多层神经网络的快速终端滑模控制策略。为了提高航空作动器响应速度和跟踪精度，设计了快速终端滑模控制策略，不仅可以加快系统响应而且可以在无扰动情况下实现系统的有限时间稳定。针对系统参数不确定性和外部扰动，设计多层神经网络进行估计并通过前馈方法加以补偿。针对神经网络的重构误差，设计了非线性鲁棒项加以克服。利用李亚普洛夫稳定性定理证明了控制系统在有扰动情况下可以实现有界稳定。实验结果表明：所设计的控制器具有良好的参数自适应和抗干扰能力，同时具有更高的跟踪精度和更快的响应速度。
Electromechanical actuators are increasingly favored by the aviation industry because of their advantages such as cleanness and easy maintenance, and the aviation electromechanical actuators have characteristics of high control accuracy, high stability and fast response speed. To further improve the response speed and tracking accuracy of aviation actuators, this paper proposes a fast terminal sliding mode control strategy based on multi-layer neural network, which can speed up the system response and realize the system limited time stability without disturbances. Aiming at the uncertainty of system parameters and external disturbances, we design a multi-layer neural network for estimation and compensation by the feedforward method. A nonlinear robust term is designed to overcome the reconstruction error of the neural network. The Lyapunov stability theorem proves that the control system can achieve bounded stability under disturbances. The experimental results show good parameter adaptation and anti-jamming capabilities, higher tracking accuracy and faster response speed of the designed controller.
[1] 杨仕鹏. 压电作动器驱动的无人直升机机体振动主动控制[D]. 南京:南京航空航天大学, 2019:50-54. YANG S P. Active vibration control of unmanned helicopter body driven by piezo actuator[D]. Nanjing:Nanjing University of Aeronautics and Astronautics, 2019:50-54(in Chinese).
[2] ZHAO Z, HE W, YIN Z, et al. Spatial trajectory tracking control of a fully actuated helicopter in known static environment[J]. Journal of Intelligent & Robotic Systems, 2016, 85(1):1-18.
[3] 岳欣, 姚建勇. 基于积分鲁棒的电液负载模拟器渐近跟踪控制[J].航空学报, 2017, 38(2):294-303. YUE X, YAO J Y. Asymptotic tracking control of electro-hydraulic load simulator based on integral robust[J]. Acta Aeronautica et Astronautica Sinica, 2017, 38(2):294-303(in Chinese).
[4] MERHRA R, CHINDE V, SINGH N M, et al. Blended algorithm for feedback linearization of single input control system[C]//IEEE International Conference on Control & Automation. Piscataway:IEEE Press, 2013.
[5] YAO B, TOMIZUKA M. Adaptive robust control of MIMO nonlinear systems in semi-strict feedback forms[J]. Automatica, 2001, 37(9):1305-1321.
[6] SONG M C,YAN L. Robust adaptive output-feedback dynamic surface control of a class of nonlinear systems with unmodeled dynamics[J]. Asian Journal of Control, 2013, 15(5):743-759.
[7] FU M Y, LIU T, LIU J, et al. Neural network-based adaptive fast terminal sliding mode control for a class of SISO uncertain nonlinear systems[C]//2016 IEEE International Conference on Mechatronics and Automation. Piscataway:IEEE Press, 2016.
[8] YU X H, ZHIHONG M. Fast terminal sliding-mode control design for nonlinear dynamical systems[J]. IEEE Transactions on Circuits & Systems I:Fundamental Theory & Applications, 2002, 49(2):261-264.
[9] 李晓刚, 方一鸣, 刘乐,等. 基于非奇异快速终端滑模的轧机液压伺服位置系统反步控制[J]. 控制与决策, 2018, 33(9):1649-1656. LI X G, FANG Y M, LIU L, et al. Backstep control of hydraulic servo position system of rolling mill based on nonsingular fast terminal sliding mode[J]. Control and Decision, 2018, 33(9):1649-1656(in Chinese).
[10] QING W, SAIF M. Robust fault detection and diagnosis in a class of nonlinear systems using a neural sliding mode observer[J]. International Journal of Systems Science, 2007, 38(11):881-899.
[11] FAN Q Y, YANG G H. Adaptive actor-critic design-based integral sliding-mode control for partially unknown nonlinear systems with input disturbances[J]. IEEE Transactions on Neural Networks & Learning Systems, 2017, 27(1):165-177.
[12] ZAK M. Terminal attractors for addressable memory in neural networks[J]. Physics Letters A, 1988, 133(1-2):18-22.
[13] HORNIK K, STINCHCOMBE M, WHITE H. Multilayer feedforward networks are universal approximators[J]. Neural Networks, 1989, 2(5):359-366.
[14] THOMAS K, SCHNETTER P, SACHS F, et al. A combined scheme of nonlinear adaptive flight control and neural network fault detection[C]//Infotech Aerospace. Reston:AIAA, 2012.
[15] CHEN H C, WANG Y H, ZHAO P. Adaptive control design for a class of nonlinear systems based on BP neural networks with scalers and saturators[C]//2016 Chinese Control and Decision Conference (CCDC). Piscataway:IEEE Press, 2016.
[16] LI B, RUI X T. Vibration control of uncertain multiple launch rocket system using radial basis function neural network[J]. Mechanical Systems and Signal Processing, 2018, 98(1):702-721.
[17] 田瑶瑶. 基于机器学习的机电系统关键部件PHM技术研究[D]. 南京:南京航空航天大学,2018:26-28. TIAN Y Y. Research on PHM technology of key components of electromechanical system based on machine learning[D]. Nanjing:Nanjing University of Aeronautics and Astronautics, 2018:26-28(in Chinese).
[18] 姜博严. 二阶系统有限时间控制问题研究[D]. 哈尔滨:哈尔滨工业大学, 2018:40-45. JIANG B Y. Research on finite-time control of second-order system[D]. Harbin:Harbin Institute of Technology, 2018:40-45(in Chinese).
[19] RIEL T, GALFFY A, SCHITTER G. Analysis and robust control of an precision motion platform using disturbance compensation[C]//2018 IEEE Conference on Control Technology and Applications (CCTA). Piscataway:IEEE Press, 2018.
[20] PARK J H, KIM S H, PARK T S. Output-feedback adaptive neural controller for uncertain pure-feedback nonlinear systems using a high-order sliding mode observer[J]. IEEE Transactions on Neural Networks and Learning Systems, 2019, 30(5):1596-1601.
[21] HU J, LIU L, WANG Y G, et al. Precision motion control of a small launching platform with disturbance compensation using neural networks[J]. International Journal of Adaptive Control & Signal Processing, 2017, 31(7):971-984.
CJA