Control Allocation of X-form Rudder Based on Neural Networks

At present, there are mainly two common frameworks for implementing the control allocation of X-form rudder: namely the pseudo-inverse framework and the optimization based framework. In order to obtain better allocation performance than the former and lower computational complexity than the latter at the same time, this paper proposes a new framework, namely the neural networks framework, for control allocation of X-form rudder. In this paper, neural networks are applied to fit the mapping from desired moments (roll moment, pitch moment and yaw moment) to the proper rudder angles. The experimental results show that the proposed method is effective and can fulfill the task of control allocation.