Adaptive neural output feedback control for stochastic nonlinear systems with unknown control directions

This paper focuses on the problem of adaptive output feedback stabilization for a class of stochastic nonlinear system with unknown control directions. By using a linear state transformation, the unknown control coefficients are lumped together, such that the original system is transformed to a new system for which control design becomes feasible. By employing the input-driven observer, a novel adaptive neural network (NN) output-feedback controller which only contains one adaptive parameter is developed for such systems by using backstepping technique and NNs´ parameterization. The proposed control design guarantees that all the signals in the closed-loop systems are 4-moment semi-globally uniformly ultimately bounded.