Control of nonlinear dynamic systems using a stability based neural network approach

A stability based approach is introduced to design neural controllers for nonlinear systems. The requisite control input is generated as the output of a neural network which is trained off-line such that the time derivative of a positive definite function of the state variables becomes negative at all points. By using the successfully trained network as a controller, it is shown that the closed-loop system can be made asymptotically stable. The stability framework introduced is shown to permit the generation of more efficient algorithms that can lead to a larger region of stability for a wide class of nonlinear systems.