A Nonlinear Neurocontrol Scheme for Spaceships

A neurocontrol scheme for the control of spaceship is proposed which associates nonlinear inversion with optimal state feedback. It was successfully applied to the lunar soft landing problem. The scheme mainly consists of two parts: First, the nonlinear dynamic inversion of the controlled object is modeled with an artificial neural network, and the controlled object is linearized by the neural inversion model. Second, based on the linearized system another artificial neural network is used as a feedback state controller to realize certain optimal control law. A simulation on computer is performed for the lunar soft landing problem. The simulation results are encouraging and show that neurocomputation could play an important role in the control of the future spaceship.