Neural Networks Based Formation Control of Anti-ship Missiles with Constant Velocity

On the basis of nonlinear model predictive control, a formation controller is designed for a group of missiles by regulating the flight path angles in a proper manner. According to the formation requirement and the practical location of the leader missile, a dynamic virtual guidance point is generated for each follower missile. The formation control of follower missiles is considered as a dynamic optimization problem. The performance index is presented in a quadratic form. The optimization problem is reformulated as a static one. The active set method is used to obtain the optimal solution in terms of three key factors offline. Therefore, a solution table is established and used to train the neural network guidance model. The online control is achieved using the real-time state measurements as model inputs. Simulation results show the effectiveness of the proposed method.