Artificial Intelligence-Assisted Spacecraft Swarm Reconfiguration Planning

Nowadays, the spacecraft swarm reconfiguration planning method of large-scale spacecraft generally faces the disadvantages of complex constraints and long calculation time, which limits the scale of spacecraft swarm in practical applications. In order to reduce the time of spacecraft swarm reconfiguration planning, this paper proposes an artificial intelligence-assisted rapid swarm reconfiguration planning method: the method first uses the Latin Hypercube Sampling method to obtain sample points and trains the BP neural network. Based on the pre-trained neural network, one can estimate the minimum distance between each spacecraft and identify the interval where collisions may occur. Finally, the convex optimization method is used to solve the swarm reconfiguration problem. The effectiveness of the method is verified by solving the 100- satellite formation reconstruction problem. The simulation results show that the number of collision constraints is reduced from 4950 to 155, and the solution time of the planning problem is about 20 s. Numerical simulation proves that the method proposed in this paper can effectively reduce the dimensionality of the large-scale cluster spacecraft reconstruction planning problem and has the potential for practical application.