Методы искусственного интеллекта для задач управления робототехническими и мехатронными системами: обзор

Развитие робототехники и мехатроники ставит перед инженерами новые задачи, привлекая для их решения методы искусственного интеллекта. Закономерно, что сложные задачи управления требуют современных решений. Дан обзор последних разработок по управлению мехатронными системами. Показана связь методов классической теории автоматического управления и машинного обучения. Кратко описаны такие известные методы классической теории управления, как оптимизация, адаптация и нечеткая логика, на основе которых построены искусственные нейронные сети и обучение с подкреплением. Рассмотрены последние достижения по применению интеллектуального управления для актуальных задач в различных областях техники. Анализ литературы показал, что будущие исследования направлены на все большую степень автоматизации и автономии объектов управления, а их свойства и характер функционирования должны приблизиться к человеческим очертаниям интеллекта.

Литература

[1] Vepa R. A review of techniques for machine learning of real-time control strategies. Intell. Syst. Eng., 1993, vol. 2, no. 2, pp. 77–90, doi: https://doi.org/10.1049/ise.1993.0009

[2] Mitchell T. Machine learning. McGraw-Hill, 1997.

[3] Mosheiov G. Scheduling problems with a learning effect. Eur. J. Oper. Res., 2001, vol. 132, no. 3, pp. 687–693, doi: https://doi.org/10.1016/S0377-2217(00)00175-2

[4] Gurel S., Akturk M. Scheduling preventive maintenance on a single CNC machine. Int. J. Prod. Res., 2008, vol. 46, no. 24, pp. 6797–6821, https://doi.org/10.1080/00207540701487833

[5] Dittrich M.-A., Uhlich F., Denkena B. Self-optimizing tool path generation for 5-axis machining processes. CIRP J. Manuf. Sci. Technol., 2019, vol. 24, pp. 49–54, doi: https://doi.org/10.1016/j.cirpj.2018.11.005

[6] Фомин В.Н., Фрадков А.Л., Якубович В.А. Адаптивное управление динамическими объектами. Москва, Наука, 1981. 447 c.

[7] Annaswamy A.M., Fradkov A.L. A historical perspective of adaptive control and learning. Annu. Rev. Control, 2021, vol. 52, pp. 18–41, doi: https://doi.org/10.1016/j.arcontrol.2021.10.014

[8] Перельман И.И. Анализ современных методов адаптивного управления с позиций приложения к автоматизации технологических процессов. Автоматика и телемеханика, 1991, № 7, с. 3–32.

[9] Андриевский Б.Р., Фрадков А.Л. Метод скоростного градиента и его приложения. Автоматика и телемеханика, 2021, № 9, с. 3–72, doi: https://doi.org/10.31857/S0005231021090014

[10] Поляк Б.Т. Введение в оптимизацию. Москва, Наука, 1983. 384 c.

[11] Карпенко А.П. Современные алгоритмы поисковой оптимизации. Москва, Изд-во МГТУ им. Н.Э. Баумана, 2014. 446 c.

[12] Николенко С., Кадурин А., Архангельская Е. Глубокое обучение. Погружение в мир нейронных сетей. Санкт-Петербург, Питер, 2021. 476 с.

[13] Заде Л. Понятие лингвистической переменной и его применение к принятию приближенных решений. Москва, Мир, 1976. 165 c.

[14] Новак В. Математические принципы нечеткой логики. Москва, Физматлит, 2006. 347 c.

[15] Cybenko G. Approximation by superpositions of a sigmoidal function. Math. Control Signal. Systems, 1989, vol. 2, no. 4, pp. 303–314, doi: https://doi.org/10.1007/BF02551274

[16] Lee J., Shin J., Realff M. Machine learning: overview of the recent progresses and implications for the process systems engineering field. Comput. Chem. Eng., 2018, vol. 114, pp. 111–121, doi: https://doi.org/10.1016/j.compchemeng.2017.10.008

[17] Sutton R.S., Barto A.G. Reinforcement learning. MIT Press, 2018. 552 p.

[18] Buoniu L., Babuska R., De Schutter B. et al. Reinforcement learning and dynamic programming using function approximators. CRC Press, 2010. 280 p.

[19] Kurrek P., Jocas M., Zoghlami F. et al. AI motion control — a generic approach to develop control policies for robotic manipulation tasks. Proc. Int. Conf. Engineering Design, 2019, vol. 1, no. 1, pp. 3561–3570, doi: https://doi.org/10.1017/dsi.2019.363

[20] Menghal P., Laxmi A. Real time simulation: a novel approach in engineering education. ICECT, 2011, vol. 1, pp. 215–219, doi: https://doi.org/10.1109/ICECTECH.2011.5941592

[21] Powell W. Approximate dynamic programming. Wiley, 2011. 656 p.

[22] Bertsekas D., Tsitsiklis J. Neuro-dynamic programming: an overview. Proc. 34th IEEE Conf. on Decision and Control, 1995, vol. 1, pp. 560–564, doi: https://doi.org/10.1109/CDC.1995.478953

[23] Prokhorov D., Wunsch D. Adaptive critic designs. IEEE Trans. Neural Netw., 1997, vol. 8, no. 5, pp. 997–1007, doi: https://doi.org/10.1109/72.623201

[24] Werbos P.J. Reinforcement learning and approximate dynamic programming (RLADP) — foundations, common misconceptions, and the challenges ahead. In: Reinforcement learning and approximate dynamic programming for feedback control. Wiley, 2012, pp. 1–30.

[25] Lewis F., Vrabie D. Reinforcement learning and adaptive dynamic programming for feedback control. IEEE Circuits Syst. Mag., 2009, vol. 9, no. 3, pp. 32–50, doi: https://doi.org/10.1109/MCAS.2009.933854

[26] Lopes G., Ferreira M., Da Silva Simoes A. et al. Intelligent control of a quadrotor with proximal policy optimization reinforcement learning. Proc. LARS/SBR/WRE, 2018, pp. 503–508, doi: https://doi.org/10.1109/LARS/SBR/WRE.2018.00094

[27] Kim J., Hong M., Lee K. et al. Learning to walk a tripod mobile robot using nonlinear soft vibration actuators with entropy adaptive reinforcement learning. IEEE Robot. Autom. Lett., 2020, vol. 5, no. 2, pp. 2317–2324, doi: https://doi.org/10.1109/LRA.2020.2970945

[28] Puccetti L., Kopf F., Rathgeber C. et al. Speed tracking control using online reinforcement learning in a real car. ICCAR, 2020, pp. 392–399, doi: https://doi.org/10.1109/ICCAR49639.2020.9108051

[29] Sun L., You F. Machine learning and data-driven techniques for the control of smart power generation systems: an uncertainty handling perspective. Engineering, 2021, vol. 7, no. 9, pp. 1239–1247, doi: https://doi.org/10.1016/j.eng.2021.04.020

[30] Tomin N., Kurbatsky V., Guliyev H. Intelligent control of a wind turbine based on reinforcement learning. ELMA, 2019, doi: https://doi.org/10.1109/ELMA.2019.8771645

[31] Sierra-Garcia J.E., Santos M., Pandit R. Wind turbine pitch reinforcement learning control improved by PID regulator and learning observer. Eng. Appl. Artif. Intell., 2022, vol. 111, art. 104769, doi: https://doi.org/10.1016/j.engappai.2022.104769

[32] Gunther J., Pilarski P., Helfrich G. et al. Intelligent laser welding through representation, prediction, and control learning: an architecture with deep neural networks and reinforcement learning. Mechatronics, 2016, vol. 34, pp. 1–11, doi: https://doi.org/10.1016/j.mechatronics.2015.09.004

[33] Lin W.-S., Yang P.-C. Adaptive critic motion control design of autonomous wheeled mobile robot by dual heuristic programming. Automatica, 2008, vol. 44, no. 11, pp. 2716–2723, doi: https://doi.org/10.1016/j.automatica.2008.03.029

[34] Calise A., Rysdyk R. Nonlinear adaptive flight control using neural networks. IEEE Control Syst. Mag., 1998, vol. 18, no. 6, pp. 14–25, doi: https://doi.org/10.1109/37.736008

[35] Kruger T., Schnetter P., Placzek R. et al. Fault-tolerant nonlinear adaptive flight control using sliding mode online learning. Neural Netw., 2012, vol. 32, pp. 267–274, doi: https://doi.org/10.1016/j.neunet.2012.02.025

[36] Lin F.-J., Hung Y.-C., Chen S.-Y. Field-programmable gate array-based intelligent dynamic sliding-mode control using recurrent wavelet neural network for linear ultrasonic motor. IET Control. Theory Appl., 2010, vol. 4, no. 9, pp. 1511–1532, doi: https://doi.org/10.1049/iet-cta.2009.0066

[37] Zhu D., Mei T., Luo M. Adaptive sliding mode control for robots based on fuzzy support vector machines. ICMA, 2009, pp. 3469–3474, doi: https://doi.org/10.1109/ICMA.2009.5246264

[38] Perez J., Cabrera J., Castillo J. et al. Bio-inspired spiking neural network for nonlinear systems control. Neural Netw., 2018, vol. 104, pp. 15–25, doi: https://doi.org/10.1016/j.neunet.2018.04.002

[39] Perez-Pena F., Lenero-Bardallo J., Linares-Barranco A. et al. Towards bioinspired close-loop local motor control: a simulated approach supporting neuromorphic implementations. IEEE ISCAS, 2017, doi: https://doi.org/10.1109/ISCAS.2017.8050808

[40] Keshta H., Ali A., Saied E. et al. Application of Static Var Compensator (SVC) with PI controller for grid integration of wind farm using harmony search. Int. J. Emerg. Electr. Power Syst., 2016, vol. 17, no. 5, pp. 555–566, doi: https://doi.org/10.1515/ijeeps-2016-0159

[41] Atashpaz Gargari E., Hashemzadeh F., Rajabioun R. et al. Colonial competitive algorithm: a novel approach for PID controller design in MIMO distillation column process. Int. J. Intell. Comput. Cybern., 2008, vol. 1, no. 3, pp. 337–355, doi: https://doi.org/10.1108/17563780810893446

[42] Schillinger M., Hartmann B., Skalecki P. et al. Safe active learning and safe Bayesian optimization for tuning a PI-controller. IFAC-PapersOnLine, 2017, vol. 50, no. 1, pp. 5967–5972, doi: https://doi.org/10.1016/j.ifacol.2017.08.1258

[43] Berkenkamp F., Schoellig A., Krause A. Safe controller optimization for quadrotors with Gaussian processes. IEEE ICRA, 2016, pp. 491–496, doi: https://doi.org/10.1109/ICRA.2016.7487170

[44] Malikopoulos A., Papalambros P., Assanis D. A learning algorithm for optimal internal combustion engine calibration in real time. Proc. DETC2007, 2008, vol. 6-A, pp. 91–100, doi: https://doi.org/10.1115/DETC2007-34718

[45] Liang D., Ding W. Modelling and predicting of a switched reluctance motor drive using radial basis function network-based adaptive fuzzy system. IET Electr. Power Appl., 2009, vol. 3, no. 3, pp. 218–230, doi: https://doi.org/10.1049/iet-epa.2008.0096

[46] Elsodany N., Rezeka S., Maharem N. Adaptive PID control of a stepper motor driving a flexible rotor. Alex. Eng. J., 2011, vol. 50, no. 2, pp. 127–136, doi: https://doi.org/10.1016/j.aej.2010.08.002

[47] Mitra P., Maulik S., Chowdhury S. et al. Anfis based automatic voltage regulator with hybrid learning algorithm. Proc. Universities Power Engineering Conf., 2007, pp. 397–401, doi: https://doi.org/10.1109/UPEC.2007.4468980

[48] Chen H.-C. Optimal fuzzy pid controller design of an active magnetic bearing system based on adaptive genetic algorithms. Proc. ICMLC, 2008, vol. 4, pp. 2054–2060, doi: https://doi.org/10.1109/ICMLC.2008.4620744

[49] Wen S., Hu X., Lv X. et al. Q-learning trajectory planning based on Takagi–Sugeno fuzzy parallel distributed compensation structure of humanoid manipulator. Int. J. Adv. Robot. Syst., 2019, vol. 16, no. 1, doi: https://doi.org/10.1177/1729881419830204

[50] Ercin O., Coban R. Identification of linear dynamic systems using the artificial bee colony algorithm. Turk. J. Elec. Eng. & Comp. Sci., 2012, vol. 20, no. 7, pp. 1175–1188, doi: https://doi.org/10.3906/elk-1012-956

[51] Agand P., Shoorehdeli M.A., Khaki-Sedigh A. Adaptive recurrent neural network with Lyapunov stability learning rules for robot dynamic terms identification. Eng. Appl. Artif. Intell., 2017, vol. 65, pp. 1–11, doi: https://doi.org/10.1016/j.engappai.2017.07.009

[52] Соболь И.М., Статников Р.Б. Выбор оптимальных параметров в задачах со многими критериями. Москва, Дрофа, 2006. 175 с.

[53] Черных В.В., Статников Р.Б., Храпов Д.С. и др. Метод исследования пространства параметров в задачах проектирования легковых автомобилей. Проблемы машиностроения и надежности машин, 2009, № 4, с. 19–25.

[54] Balochian S., Baloochian H. Social mimic optimization algorithm and engineering applications. Expert Syst. Appl., 2019, vol. 134, pp. 178–191, doi: https://doi.org/10.1016/j.eswa.2019.05.035

[55] Kumar S., Kumar A., Shankar G. Crow search algorithm based optimal dynamic performance control of SVC assisted SMIB system. NPSC, 2018, doi: https://doi.org/10.1109/NPSC.2018.8771814

[56] Abualigah L. Multi-verse optimizer algorithm: a comprehensive survey of its results, variants, and applications. Neural Comput. & Applic., 2020, vol. 32, no. 16, pp. 12381–12401, doi: https://doi.org/10.1007/s00521-020-04839-1

[57] Wang L., Liu Z., Chen C. et al. Support vector machine based optimal control for minimizing energy consumption of biped walking motions. Int. J. Precis. Eng. Manuf., 2012, vol. 13, no. 11, pp. 1975–1981, doi: https://doi.org/10.1007/s12541-012-0260-7

[58] Xiao L., Zhu Q., Li C. et al. Application of modified teaching-learning algorithm in coordination optimization of TCSC and SVC. In: CCPR. Springer, 2014, pp. 44–53, doi: https://doi.org/10.1007/978-3-662-45646-0_5

[59] Shouran M., Habil M. Tuning of PID controller using different optimization algorithms for industrial DC motor. ICACITE, 2021, pp. 756–759, doi: https://doi.org/10.1109/ICACITE51222.2021.9404616

[60] Silvestri M., Pedrazzoli P., Boer C. et al. Compensating high precision positioning machine tools by a self learning capable controller. Proc. EUSPEN, 2011, vol. 2, pp. 121–124.

[61] Tomizuka M. Zero phase error tracking algorithm for digital control. J. Dyn. Sys., Meas., Control., 1987, vol. 109, no. 1, pp. 65–68, doi: https://doi.org/10.1115/1.3143822

[62] Tomizuka M. Dealing with periodic disturbances in controls of mechanical systems. Annu. Rev. Control, 2008, vol. 32, no. 2, pp. 193–199, doi: https://doi.org/10.1016/j.arcontrol.2008.07.002

[63] Uchiyama M. Formation of high-speed motion pattern of a mechanical arm by trial. Transactions of the Society of Instrument and Control Engineers, 1978, vol. 14, no. 6, pp. 706–712, doi: https://doi.org/10.9746/sicetr1965.14.706

[64] Li W., Maier P., Enge H. Self-learning control applied to vibration control of a rotating spindle by piezopusher bearings. Proc. Inst. Mech. Eng., I: J. Syst. Control Eng., 2004, vol. 218, no. 3, pp. 185–196, doi: https://doi.org/10.1243/095965104773913867

[65] Hladowski L., Galkowski K., Rogers E. Further results on dynamic iterative learning control law design using repetitive process stability theory. 10th nDS, 2017, doi: https://doi.org/10.1109/NDS.2017.8070621

[66] Kurek J., Zaremba M. Iterative learning control synthesis based on 2-D system theory. IEEE Trans. Automat. Contr., 1993, vol. 38, no. 1, pp. 121–125, doi: https://doi.org/10.1109/9.186321

[67] Pakshin P., Emelianova J., Emelianov M. Iterative learning control design for multiagent systems based on 2D models. Autom. Remote Control, 2018, vol. 79, no. 6, pp. 1040–1056, https://doi.org/10.1134/S000511791806005X

[68] Seel T., Werner C., Raisch J. et al. Iterative learning control of a drop foot neuroprosthesis — generating physiological foot motion in paretic gait by automatic feedback control. Control. Eng. Pract., 2016, vol. 48, pp. 87–97, doi: https://doi.org/10.1016/j.conengprac.2015.11.007

[69] Seel T., Werner C., Schauer T. The adaptive drop foot stimulator — multivariable learning control of foot pitch and roll motion in paretic gait. Med. Eng. Phys., 2016, vol. 38, no. 11, pp. 1205–1213, doi: https://doi.org/10.1016/j.medengphy.2016.06.009

[70] Mohammadpour A., Mishra S., Parsa L. Iterative learning control for fault-tolerance in multi-phase permanent-magnet machines. Proc. American Control Conf., 2013, pp. 5929–5934, doi: https://doi.org/10.1109/acc.2013.6580768

[71] Li J., Wang S., Wang J. et al. Iterative learning control for a distributed cloud robot with payload delivery. Assem. Autom., 2021, vol. 41, no. 3, pp. 263–273, doi: https://doi.org/10.1108/AA-11-2020-0179

[72] Warrier R., Devasia S. Iterative learning from novice human demonstrations for output tracking. Trans. Hum. Mach. Syst., 2016, vol. 46, no. 4, pp. 510–521, doi: https://doi.org/10.1109/THMS.2016.2545243

[73] Lansdell B., Kording K. Towards learning-to-learn. Curr. Opin. Behav. Sci., 2019, vol. 29, pp. 45–50, doi: https://doi.org/10.1016/j.cobeha.2019.04.005