Stable adaptive neuro-control design via Lyapunov function derivative estimation (Q5943840)

A new adaptive technique for solving the tracking problem for affine control nonlinear dynamical systems, based on estimating the derivative of an unknown Lyapunov function and exploiting the approximation capabilities of the neural networks structures that are linear in the weights, is presented. Besides, such a controller incorporates the novel resetting strategy to avoid possible division by zero. The developed control scheme guarantees the uniform ultimate boundedness of the tracking error with respect to an arbitrarily small set plus the boundedness of all other signals in the closed loop. To this end, the proposed scheme is tested via simulations performed on a simple example. Lyapunov function like techniques have long been used in studying control problems for nonlinear systems. But unfortunately, even for completely known systems, the construction of a Lyapunov function is in general a challenging task. So, obviously, the problem becomes intractable for unknown systems, too. So this flaw cannot be removed.