Stability, convergence, and verification and validation challenges of neural net adaptive flight control

Abstract

This paper provides a discussion of challenges of neural net adaptive flight control and an examination of stability and convergence issues of adaptive control algorithms. Understanding stability and convergence issues with adaptive control is important in order to advance adaptive control to a higher technology readiness level. The stability and convergence of neural net learning law are investigated. The effect of unmodeled dynamics on learning law is examined. Potential improvements in the learning law and adaptive control architecture based on optimal estimation are presented. The paper provides a brief summary of the future research of the Integrated Resilient Aircraft Control (IRAC) in the area of adaptive flight control. The paper also discusses challenges and future research in verification and validation. © 2010 Springer-Verlag Berlin Heidelberg.