State Feedback Stabilization of Nonlinear Systems via the Neural Network Approach

We consider the state feedback stabilization of autonomous nonlinear systems described by dx/dt = Ax + Bu - f(x), where f(x) is a memoryless nonlinearity and does not necessarily satisfy the sector conditions. Classical results can not be used to infer stability of the closed loop system. By using neural network techniques, however, we find a state feedback gain matrix that ensures the asymptotic stabilit for any specified equilibrium.