| Peer-Reviewed

FPGA Implementation of Neural Network-Based AGPC for Nonlinear F-16 Aircraft Auto-pilot Control: Part 1 – Modeling, Synthesis, Verification and FPGA-in-Loop Co-Sim

Received: 24 April 2022     Accepted: 9 June 2022     Published: 5 September 2022
Views:       Downloads:
Abstract

Model predictive control (MPC) is an advanced receding horizon control strategy, for difficult multivariable control systems, that leads to constrained optimization problems which are solved online at each sampling time interval, and takes full advantage of the computational power available in modern control computer hardware for hard real-time constraint systems with short sampling time. However, nonlinear MPC (NMPC) attracts additional computational overload to satisfy nonlinear systems with hard real-time constraint and relatively short sampling time. In other to deploy NMPC for the control of nonlinear systems with hard real-time constraint and relatively short sampling time, a new model-based design (MBD) approach for the implementation of nonlinear MPC, called adaptive general predictive control (AGPC), on field programmable gate array (FPGA) for the auto-pilot control of a nonlinear F-16 aircraft is presented in this paper. The new MBD approach consists of four parts: (i) In the model identification part, the nonlinear F-16 aircraft model is approximated by a neural network autoregressive moving average with exogenous inputs (NNARMAX) model which is trained by an adaptive recursive least squares (ARLS) algorithm; (ii) In the adaptive control part, the nonlinear F-16 aircraft is controlled by a constrained neural network-based adaptive generalized predictive control (AGPC) algorithm; (iii) The third part is the online closed-loop NNARMAX model identification and AGPC control of the nonlinear F-16 aircraft; and (iv) The modeling, synthesis, verification and FPGA-in-the-loop hardware co-simulation (HW Co-Sim) of the online closed-loop NNARMAX model identification and AGPC control of the nonlinear F-16 aircraft. The training and validation data for the neural network model identification are obtained from the open-loop simulation of first-principle nonlinear F-16 aircraft model. The online closed-loop NNARMAX model identification and AGPC control of the nonlinear F-16 aircraft demonstrates the efficiency of the ARLS and the AGPC algorithms in tracking the desired reference trajectories of the nonlinear F-16 aircraft. The FPGA-in-the-loop hardware co-simulation of the online closed-loop NNARMAX model identification and AGPC control of the nonlinear F-16 aircraft shows significant reduction in the computation time between the floating-point MATLAB AGPC and fixed-point C++ AGPC algorithms. Hence, the effort in this work has been directed towards reducing the computation time of the AGPC algorithm at each sampling time instant through modeling, synthesizing and mapping the AGPC algorithm to Virtex-5 FX70T ML507 FPGA embedded system development board via FPGA-in-the-loop hardware co-simulation verification which has been successfully achieved and validated.

Published in American Journal of Embedded Systems and Applications (Volume 9, Issue 1)
DOI 10.11648/j.ajesa.20220901.13
Page(s) 6-36
Creative Commons

This is an Open Access article, distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution and reproduction in any medium or format, provided the original work is properly cited.

Copyright

Copyright © The Author(s), 2022. Published by Science Publishing Group

Keywords

Adaptive Generalized Predictive Control, Adaptive Recursive Least Squares, Auto-pilot Control, Hardware Co-simulation, Nonlinear Model-Based Design, NMPC, Neural Networks, Nonlinear F-16 Aircraft

References
[1] R. S. Russel, “Non-linear F-16 Simulation using Simulink and Matlab, Ver. 1.0”, Technical Report, University of Minnesota, United States of America, 2003.
[2] B. L. Stevens and F. L. Lewis, “Aircraft Control and Simulation”. 2nd ed., New York: John Wiley & Sons, Inc., 2003.
[3] V. A. Akpan and G. D. Hassapis, “Training dynamic feedforward neural networks for online nonlinear model identification and control applications”. International Reviews of Automatic Control: Theory & Applications, vol. 4, no. 3, pp. 335-350, 2011.
[4] V. A. Akpan and G. D. Hassapis, “Nonlinear model identification and adaptive model predictive control using neural networks”, ISA Transactions; vol. 5, no. 2, pp. 177–94, 2011.
[5] V. A. Akpan, “Development of new model adaptive predictive control algorithms and their implementation on real-time embedded systems”, Aristotle university of Thessaloniki, GR-54124, Thessaloniki, Greece, Ph.D. Dissertation, 517 pages, July, 2011. Available [Online]: http://invenio.lib.auth.gr/record/127274/files/GRI-2011-7292.pdf.
[6] L. T. Nguyen, M. E. Ogburn, W. P. Gilbert, K. S. Kibler, P. W. Brown and P. L. Deal, “Simulator Study of Stall/Post-Stall Characteristics of a Fighter Airplane With Relaxed Longitudinal Static Stability”. NASA Technical Paper 1538, Dec., 1979, 233 pages.
[7] A. Savran, R. Tasaltin and Y. Becerikli. “Intelligent adaptive nonlinear flight control for a high performance aircraft with neural networks”. ISA Transactions, vol. 45, no. 2, pp. 225-247, 2006.
[8] M. Hanhila, T. Mantere and J. T. Alander, “FPGA–implementation of PID-controller by differential evolution optimization”, DE GRUYTER: Open Engineering, vol. 8, pp. 395-402, 2017.
[9] J. M. Maciejowski, “Predictive Control with Constraints”. England: Pearson Education Limited, 2002.
[10] E. F. Camacho and C. Bordons, “Model Predictive Control”. 2nd ed., London: Springer-Verlag, 2007.
[11] J. E. Normey-Rico and E. F. Camacho, “Control of Dead-Time Processes”. London: Springer-Verlag, 2007.
[12] K. V. Ling, B. F. Wu and J. M. Maciejowski, “Embedded model predictive control (MPC) using a FPGA”. In Proceedings of the 17th World Congress, The International Federation of Automatic Control (IFAC), Seoul, Korea, July 6-11, 2008, pp. 15250-15255, 2008.
[13] E. Hartley, “Model predictive control on an FPGA: Aerospace and Space Scenarios”, Workshop on Embedded Optimization EMBOPT 2014, IMT Lucca, University of Cambridge, United Kingdom, 8th September, 2014, pp. 1-53. Available [Online]: https://www.imtlucca.it/embopt-14/Slides/hartley.pdf.
[14] Y. Kondratenko and E. Gordienko, “Implementation of the neural networks for adaptive control system on FPGA” Annals of DAAAM for 2012 & Proceedings of the 23rd International DAAAM Symposium, Vienna-Austria, vol. 23, no. 1, pp. 1-4, 2012.
[15] S. Gerkšič and B. Pregelj, “Finite-word-length FPGA implementation of model predictive control for ITER resistive wall mode control”, Fusion Engineering and Design, vol. pp. 112480, 2021. DOI: https://doi.org/10.1016/j.fusengdes.2021.112480.
[16] K. Mohamed, A. E. Mahdy and M. Refai, “Model predictive control using FPGA”, International Journal of Control Theory and Computer Modeling, vol. 5, no. 2, pp. 1-14, 2015.
[17] S. Lucia, D. Navarro, O. Lucia, P. Zometa and R. Findeisen, “Optimized FPGA implementation of model predictive control for embedded systems using high level synthesis tool”, IEEE Transactions on Industrial Informatics, vol. 14, no. 1, pp. 137-145, 2018.
[18] Y. Shoukry, M. W. El-Kharashi and S. Hammad, “MPC-on-chip: An embedded GPC coprocessor for automotive active suspension systems”. IEEE Embedded Systems Letters, vol. 2, no. 2, pp. 31-34, 2010.
[19] P. Vouzis, L. G. Bleris, M. Arnold and M. V. Kothare, “A custom-made algorithmic-specific processor for model predictive control”. In Proceedings of the International Symposium on Industrial Electronics, Montreal, Canada, 9-13 June, 2006.
[20] E. N. Hartley, J. Jerez, A. Suardi, J. Maciejowski, E. C. Kerrigan and G. A. Constantinides, “Predictive control using an FPGA with application to aircraft control”, IEEE Transactions on Control Systems Technology, vol. 22, no. 3, pp. 1006-1017, 2014. DOI: 10.1109/TCST.2013.2271791.
[21] R. DeMott, “Development of a flexible FPGA-based platform for flight control system research”, Ph.D. Theses and Dissertation, Graduate School, Virginia Commonwealth University, Richmond, Virginia, United States of America, pp. 1-157, 2010. Available [Online]: https://core.ac.uk/download/pdf/51289440.pdf.
[22] M. Aboelaze, O. El-Debb, A. E. Mansour and M. A. Ghazy, “FPGA implementation of a satellite altitude control using variable structure control”, In Proceedings of the 2nd International Conference on Computer Science and Engineering (MIC-Computing 2014), Milan Italy, pp. 1-6, 2014. DOI: 10.13140/RG.2.1.4462.5448.
[23] H. U. Azad, D. V. Lazic and W. Shahid, “FPGA based longitudinal and lateral controller implementation for a small UAV”, World Academy of Science, Engineering and Technology, vol. 46, pp. 10-22, 2010.
[24] L. Kalra and C. Georgakis, “Effects of process nonlinearity on the performance of linear model predictive controllers for the environmentally safe operation of a fluid catalytic cracking unit”. Industrial Engineering Chemical Research, vol. 33, pp. 3063-3069, 1994.
[25] B. H. Fletcher, “FPGA embedded processors: Revealing true system performance”. Embedded Training Program, Embedded System Conference, San Francisco, USA, 6-10 Mar., 2005, ETP–367, pp. 1-18, 2005.
[26] V. A. Akpan, “Model-based embedded-processor systems design methodologies: Modeling, syntheses, implementation and validation”, African Journal of Computing and ICTs, vol. 5, no. 1, pp. 1-26, 2012.
[27] V. A. Akpan, “Hard and soft embedded FPGA processor systems design: Design considerations and performance comparisons”, International Journal of Engineering and Technology, vol. 3, no. 11, pp. 1000-1020, 2013.
[28] V. A. Akpan, “An FPGA realization of integrated embedded multi-processors system: A hardware-software co-design approach”, Journal of Advanced Research in Embedded System, vol. 2, no. 1, pp. 1-27, 2015.
[29] V. A. Akpan, “FPGA-in-the-Loop implementation of an adaptive matrix inversion algorithmic co-processor: A dual-processor system design”, Journal of Advanced Research in Embedded System, vol. 2, no. 1, pp. 28-63, 2015.
[30] H. K. Khalil, “Nonlinear Systems”. Upper Saddle River, NJ: Prentice-Hall, 1996.
[31] M. Morari and E. Zafiriou, “Robust Process Control”, Englewood Cliffs, NJ: Prentice-Hall, 1989.
[32] M. Nørgaard, O. Ravn, N. K. Poulsen and L. K. Hansen, “Neural Networks for Modelling and Control of Dynamic Systems: A Practitioner’s Handbook”, London: Springer-Verlag, 2000.
[33] D. F. Anderson and S. Eberhardt, “Understanding Flight”, New York, U.S.A.: McGraw-Hill, 2001.
[34] D. F. Anderson and S. Eberhardt, “Understanding Flight”, 2nd Ed., New York, U.S.A.: McGraw-Hill, 2010.
[35] G. J. J. Ducard, “Fault-Tolerant Flight Control and Guidance Systems: Practical Methods for Small Unmanned Aerial Vehicles”, London: Springer-Verlag Limited, 2009.
[36] R. C. Nelson, “Flight Stability and Automatic Control”, New York: McGraw-Hill, Inc, 1989.
[37] The MathWorks Inc., MATLAB® & Simulink® R2021a, Natick, USA. www.mathworks.com.
[38] Xilinx Inc., “Xilinx Software Design Suite 14.5”, (2022). USA. www.xilinx.com.
[39] P. Meloni, S. Secchi and L. Raffo, “An FPGA-based framework for technology-aware prototyping of multicore embedded architectures”. IEEE Embedded Syst. Letters, vol. 2, no. 1, pp. 5-9, 2010.
[40] Xilinx Inc., “Xilinx AccelDSP Synthesis Tool”, User Guide, v11.4, December 2, 2009, pp. 1-222.
[41] Xilinx Inc., “Xilinx System Generator for DSP”, User Guide, v12.1, April 19, 2014, pp. 1-414.
[42] Sjöberg, J and Ljung, L. (1995). “Overtraining, regularization, and searching for minimum in neural networks”. International Journal of Control, vol. 62, pp. 1391-1408.
Cite This Article
  • APA Style

    Vincent Andrew Akpan, Dimitrios Chasapis, George Dimitriou Hassapis. (2022). FPGA Implementation of Neural Network-Based AGPC for Nonlinear F-16 Aircraft Auto-pilot Control: Part 1 – Modeling, Synthesis, Verification and FPGA-in-Loop Co-Sim. American Journal of Embedded Systems and Applications, 9(1), 6-36. https://doi.org/10.11648/j.ajesa.20220901.13

    Copy | Download

    ACS Style

    Vincent Andrew Akpan; Dimitrios Chasapis; George Dimitriou Hassapis. FPGA Implementation of Neural Network-Based AGPC for Nonlinear F-16 Aircraft Auto-pilot Control: Part 1 – Modeling, Synthesis, Verification and FPGA-in-Loop Co-Sim. Am. J. Embed. Syst. Appl. 2022, 9(1), 6-36. doi: 10.11648/j.ajesa.20220901.13

    Copy | Download

    AMA Style

    Vincent Andrew Akpan, Dimitrios Chasapis, George Dimitriou Hassapis. FPGA Implementation of Neural Network-Based AGPC for Nonlinear F-16 Aircraft Auto-pilot Control: Part 1 – Modeling, Synthesis, Verification and FPGA-in-Loop Co-Sim. Am J Embed Syst Appl. 2022;9(1):6-36. doi: 10.11648/j.ajesa.20220901.13

    Copy | Download

  • @article{10.11648/j.ajesa.20220901.13,
      author = {Vincent Andrew Akpan and Dimitrios Chasapis and George Dimitriou Hassapis},
      title = {FPGA Implementation of Neural Network-Based AGPC for Nonlinear F-16 Aircraft Auto-pilot Control: Part 1 – Modeling, Synthesis, Verification and FPGA-in-Loop Co-Sim},
      journal = {American Journal of Embedded Systems and Applications},
      volume = {9},
      number = {1},
      pages = {6-36},
      doi = {10.11648/j.ajesa.20220901.13},
      url = {https://doi.org/10.11648/j.ajesa.20220901.13},
      eprint = {https://article.sciencepublishinggroup.com/pdf/10.11648.j.ajesa.20220901.13},
      abstract = {Model predictive control (MPC) is an advanced receding horizon control strategy, for difficult multivariable control systems, that leads to constrained optimization problems which are solved online at each sampling time interval, and takes full advantage of the computational power available in modern control computer hardware for hard real-time constraint systems with short sampling time. However, nonlinear MPC (NMPC) attracts additional computational overload to satisfy nonlinear systems with hard real-time constraint and relatively short sampling time. In other to deploy NMPC for the control of nonlinear systems with hard real-time constraint and relatively short sampling time, a new model-based design (MBD) approach for the implementation of nonlinear MPC, called adaptive general predictive control (AGPC), on field programmable gate array (FPGA) for the auto-pilot control of a nonlinear F-16 aircraft is presented in this paper. The new MBD approach consists of four parts: (i) In the model identification part, the nonlinear F-16 aircraft model is approximated by a neural network autoregressive moving average with exogenous inputs (NNARMAX) model which is trained by an adaptive recursive least squares (ARLS) algorithm; (ii) In the adaptive control part, the nonlinear F-16 aircraft is controlled by a constrained neural network-based adaptive generalized predictive control (AGPC) algorithm; (iii) The third part is the online closed-loop NNARMAX model identification and AGPC control of the nonlinear F-16 aircraft; and (iv) The modeling, synthesis, verification and FPGA-in-the-loop hardware co-simulation (HW Co-Sim) of the online closed-loop NNARMAX model identification and AGPC control of the nonlinear F-16 aircraft. The training and validation data for the neural network model identification are obtained from the open-loop simulation of first-principle nonlinear F-16 aircraft model. The online closed-loop NNARMAX model identification and AGPC control of the nonlinear F-16 aircraft demonstrates the efficiency of the ARLS and the AGPC algorithms in tracking the desired reference trajectories of the nonlinear F-16 aircraft. The FPGA-in-the-loop hardware co-simulation of the online closed-loop NNARMAX model identification and AGPC control of the nonlinear F-16 aircraft shows significant reduction in the computation time between the floating-point MATLAB AGPC and fixed-point C++ AGPC algorithms. Hence, the effort in this work has been directed towards reducing the computation time of the AGPC algorithm at each sampling time instant through modeling, synthesizing and mapping the AGPC algorithm to Virtex-5 FX70T ML507 FPGA embedded system development board via FPGA-in-the-loop hardware co-simulation verification which has been successfully achieved and validated.},
     year = {2022}
    }
    

    Copy | Download

  • TY  - JOUR
    T1  - FPGA Implementation of Neural Network-Based AGPC for Nonlinear F-16 Aircraft Auto-pilot Control: Part 1 – Modeling, Synthesis, Verification and FPGA-in-Loop Co-Sim
    AU  - Vincent Andrew Akpan
    AU  - Dimitrios Chasapis
    AU  - George Dimitriou Hassapis
    Y1  - 2022/09/05
    PY  - 2022
    N1  - https://doi.org/10.11648/j.ajesa.20220901.13
    DO  - 10.11648/j.ajesa.20220901.13
    T2  - American Journal of Embedded Systems and Applications
    JF  - American Journal of Embedded Systems and Applications
    JO  - American Journal of Embedded Systems and Applications
    SP  - 6
    EP  - 36
    PB  - Science Publishing Group
    SN  - 2376-6085
    UR  - https://doi.org/10.11648/j.ajesa.20220901.13
    AB  - Model predictive control (MPC) is an advanced receding horizon control strategy, for difficult multivariable control systems, that leads to constrained optimization problems which are solved online at each sampling time interval, and takes full advantage of the computational power available in modern control computer hardware for hard real-time constraint systems with short sampling time. However, nonlinear MPC (NMPC) attracts additional computational overload to satisfy nonlinear systems with hard real-time constraint and relatively short sampling time. In other to deploy NMPC for the control of nonlinear systems with hard real-time constraint and relatively short sampling time, a new model-based design (MBD) approach for the implementation of nonlinear MPC, called adaptive general predictive control (AGPC), on field programmable gate array (FPGA) for the auto-pilot control of a nonlinear F-16 aircraft is presented in this paper. The new MBD approach consists of four parts: (i) In the model identification part, the nonlinear F-16 aircraft model is approximated by a neural network autoregressive moving average with exogenous inputs (NNARMAX) model which is trained by an adaptive recursive least squares (ARLS) algorithm; (ii) In the adaptive control part, the nonlinear F-16 aircraft is controlled by a constrained neural network-based adaptive generalized predictive control (AGPC) algorithm; (iii) The third part is the online closed-loop NNARMAX model identification and AGPC control of the nonlinear F-16 aircraft; and (iv) The modeling, synthesis, verification and FPGA-in-the-loop hardware co-simulation (HW Co-Sim) of the online closed-loop NNARMAX model identification and AGPC control of the nonlinear F-16 aircraft. The training and validation data for the neural network model identification are obtained from the open-loop simulation of first-principle nonlinear F-16 aircraft model. The online closed-loop NNARMAX model identification and AGPC control of the nonlinear F-16 aircraft demonstrates the efficiency of the ARLS and the AGPC algorithms in tracking the desired reference trajectories of the nonlinear F-16 aircraft. The FPGA-in-the-loop hardware co-simulation of the online closed-loop NNARMAX model identification and AGPC control of the nonlinear F-16 aircraft shows significant reduction in the computation time between the floating-point MATLAB AGPC and fixed-point C++ AGPC algorithms. Hence, the effort in this work has been directed towards reducing the computation time of the AGPC algorithm at each sampling time instant through modeling, synthesizing and mapping the AGPC algorithm to Virtex-5 FX70T ML507 FPGA embedded system development board via FPGA-in-the-loop hardware co-simulation verification which has been successfully achieved and validated.
    VL  - 9
    IS  - 1
    ER  - 

    Copy | Download

Author Information
  • Department of Biomedical Technology, The Federal University of Technology, Akure, Nigeria

  • Barcelona Supercomputing Center, Group Sonar, Barcelona, Spain

  • Department of Electrical and Computer Engineering, Aristotle University of Thessaloniki, Thessaloniki, Greece

  • Sections