Viability Approach to Simulation of an Adaptive Controller

This paper addresses the analysis of an adaptive aircraft controller using viability theory. State variables of the controller are constrained according to admissible deviation magnitudes, and differential games where controls work against uncertainties are considered. Viability theory is used to compute viability kernels, maximal subsets of the state constraint domain where controller's trajectories can remain arbitrary long if the first player utilizes an appropriate feedback strategy, and the second player generates any admissible time-dependent uncertainties. It is shown that the controller itself is sufficiently stable so that it can successfully work against worst-case uncertainties using constant controls, i.e. constant learning rate coefficients.     «

This paper addresses the analysis of an adaptive aircraft controller using viability theory. State variables of the controller are constrained according to admissible deviation magnitudes, and differential games where controls work against uncertainties are considered. Viability theory is used to compute viability kernels, maximal subsets of the state constraint domain where controller's trajectories can remain arbitrary long if the first player utilizes an appropriate feedback strategy, and the...     »