Multi-Target Rapid Mission Planning via Lambert Transfer

A rapid planning method for a spacecraft to successively rendezvous with multiple targets via Lambert transfer is presented in this paper. For the one-to-many trajectory planning at the lower level, based on the deep neural network and the equal division of duration constraint, a highly efficient calculation module is constructed. The number of times to solve the Lambert problem corresponding to each target is reduced to 2, and there is no problem of the convergence performance degradation of the optimization algorithm due to the increase in the dimensionality of variables. On this basis, a sequence planning network is constructed based on the attention mechanism. After being trained through reinforcement learning, it can achieve the rapid planning of the optimal rendezvous sequence for a fixed number of targets. The effectiveness of the proposed multi-target mission planning method is verified through simulations. The results show that the proposed two-layer structure algorithm has significant efficiency advantages compared with the traditional iterative calculation process.