Adaptive neural network output feedback passivation of MIMO uncertain systems

An adaptive output feedback neural network controller is designed, capable of rendering affine in the control uncertain MIMO nonlinear systems passive. Consequently, a linear output feedback is employed to stabilize the system. The controlled system need not be in normal form, or have a well defined relative degree. Under a zero state detectability assumption, uniform ultimate boundedness of the system´s state is guaranteed, along with the boundedness of all other signals in the closed loop. The proposed adaptive controller is of switching type, to effectively avoid possible division by zero, while guaranteeing the continuity of switchings, thus alleviating drawbacks connected to existence of solutions and chattering phenomena. Simulations illustrate the approach.