Neural integrated control for a free-floating space robot with suddenly changing parameters

Because the state of a free-floating space robot model is uncertain and sudden changes in the model parameters might undermine the stability of the system, this paper proposes a control strategy based on a variable structure neural integrated controller. This scheme does not need a precise space robot model, making use of the radial basis function neural network ability approach to learn about an uncertain model. The network weights are adjusted online in real-time. During the early period of the control phase and parameter changes, the variable structure controller compensates for the uncertain model which the neural network could not learn well. It also creates global asymptotic stability for the whole closed-loop system. Simulation results show that the controller can handle bad changeable conditions and has important application value for defense, aerospace and other major security fields.