Output Feedback Tracking Control Based on Neural Network for a Class of SISO Strict Feedback Nonlinear Systems

The paper proposes a new output feedback tracking controller using neural network (NN) for a class of SISO strict-feedback nonlinear systems that only the output variables can be measured. The distinguished aspect of the controller is that no backstepping design is employed, and the strict-feedback systems could be transformed into the standard affine form. The gains of observer and controller are simultaneously tuned according to output tracking error based on non-separation principle design. With the universal approximation property of NN and the simultaneous parametrisation, no Lipschitz assumption and SPR condition are employed which makes the system construct simple. The proposed neural network controller can guarantee that output tracking error and all the states in the closed-loop system are the semi-globally ultimately bounded by Lyapunov approach. Finally the simulation results are used to demonstrate the effectiveness of the control scheme.