ABSTRACT
The training of the neural network is performed on-line using an extended Kalman
filter. A block transformation of the neural model to solve the inverse optimal trajec-
tory tracking as a stabilization problem for block-control-form nonlinear systems is
included. Applicability of the proposed control schemes is illustrated via simulations.
Due to favorable stability margins of optimal control systems, we synthesize a stabi-
lizing feedback control law, which will be optimal with respect to a cost functional.
At the same time, we want to avoid the difficult task of solving the associated HJB
partial differential equation. In the inverse optimal control approach, a CLF candidate
is used to construct an optimal control law directly without solving the HJB equation
[7], and still allowing us to obtain Kalman-type stability margins [14]. A storage
function is used as a CLF candidate and the inverse optimal control law is selected as
an output feedback one, which is obtained as a result of solving the Bellman equation.
Then, a CLF candidate for the obtained control law is proposed such that it stabilizes
the system and a posteriori cost functional is minimized. For this control scheme, a
