Toward V&V of neural network based controllers

ABSTRACT

Online adaptation is a powerful means to handle unexpected slow or catastrophic changes of the system's behavior (e.g., a stuck or broken rudder of an aircraft). Therefore, adaptation is one way for realizing a self-healing system. Substantial research and development has been made to use neural networks (NN) for such tasks (e.g., integrated in various unmanned helicopters and test-flown on a modified F-15 aircraft). Despite the advantages of adaptive neural network based systems, the lack of methods to perform certification, verification, and validation (V&V) of such systems severely restricts their applicability.In this paper, we report on ongoing work to develop V&V techniques and processes for NN-based safety-critical control systems, in our case an aircraft flight control system. Although the project ultimately aims at V&V of online adaptive systems, this paper focuses on the first part of this project dealing with so-called pre-trained neural networks (PTNN). V&V techniques developed here are important pre-requisites for handling the online adaptive case. In particular, we describe highlights of a process guide which has been developed within this project and discuss important V&V issues which need to be addressed during certification.

REFERENCES

• 1: Christopher M. Bishop, Neural Networks for Pattern Recognition, Oxford University Press, Inc., New York, NY, 1995
• 2: A. Calise and R. Rysdyk. Adaptive model inversion flight control for tiltrotor aircraft. In AIAA Guidance, Navigation and Control Conference 1997, number AIAA-97-3758. AIAA, 1997.
• 3: J. E. Dennis, Jr. , Robert B. Schnabel, Numerical Methods for Unconstrained Optimization and Nonlinear Equations (Classics in Applied Mathematics, 16), Soc for Industrial & Applied Math, 1996
• 4: DO-178B: Software considerations in airborne systems and equipment certification. URL: http://www.rtca.org, 1992.
• 5: Shuzhi S. Ge , Christopher J. Harris, Adaptive Neural Network Control of Robotic Manipulators, World Scientific Publishing Co., Inc., River Edge, NJ, 1998
• 6: P. Gill, W. Murray, and M. Wright. Practical Optimization. Academic Press, 1981.
• 7: M. Idan, M. Johnson, and A. Calise. A hierarchical approach to adaptive control for improved flight safety. In AIAA Guidance, Navigation and Control Conference 2001, number AIAA-2001-4209. AIAA, 2001.
• 8: IEEE standards 12207.0, 12207.1, 12207.2. URL: http://ieeexplore.ieee.org, 1997.
• 9: C. Jorgensen. Direct adaptive aircraft control using neural networks. Technical Report TM-47136, NASA, 1997.
• 10: J. Kaneshige and K. Gundy-Burlet. Integrated neural flight and propulsion control system. In AIAA Guidance, Navigation and Control Conference 2001, number AIAA-2001-4386. AIAA, 2001.
• 11: A. Kelkar. Neural networks for modeling and control of dynamic systems. Presentation at NASA Ames, Code IC, 2001.
• 12: D. Mackall, S. Nelson, and J. Schumann. Verification and Validation of Neural Networks of Aerospace Applications. Technical Report CR-211409, NASA, 2002.
• 13: NASA guidebook for safety critical software. Technical Report NASA-GB-1740.13-96, NASA, 1996.
• 14: NASA procedures and guidelines NPG: 2820.draft, NASA software guidelines and requirements as of 3/19/01. NASA Ames Research Center, Moffett Field, California, USA, 2001. (Responsible Office: Code AE/Office of the Chief Engineer).
• 15: Magnus Nrgaard , O. E. Ravn , N. K. Poulsen , L. K. Hansen, Neural Networks for Modelling and Control of Dynamic Systems: A Practitioner's Handbook, Springer-Verlag New York, Inc., Secaucus, NJ, 2000
• 16: Shankar Sastry , Marc Bodson, Adaptive control: stability, convergence, and robustness, Prentice-Hall, Inc., Upper Saddle River, NJ, 1989
• 17: J. Schumann. Vericonn: Verification of controllers based on adaptive neural networks --- white paper---. Technical report, NASA Ames, Automated Software Engineering, 2001.
• 18: J. Schumann. V&V issues for neural networks. Technical Report RIACS-TR-XX-02, RIACS, 2002.
• 19: D. Soloway and P. Haley. Reconfigurable flight control using neural generalized predictive control. In AIAA Space 2000 Conference, number AIAA-2000-5328. AIAA, 2000.
• 20: J. Totah. Adaptive flight control and on-line learning. In AIAA Guidance, Navigation and Control Conference 1997, number AIAA-97-3537. AIAA, 1997.