Optimal Satellite Formation Reconfiguration Using Pseudospectral Method and Co-evolutionary Particle Swarm Optimization

Abstract

Effective formation reconfiguration of satellite flying is increasingly becoming important. This paper presents an alternative algorithm for satellite formation reconfiguration based on the hybridization of Legendre pseudospectral method (LPM) and co-evolutionary particle swarm optimization (CPSO), which aims to give a fuel-optimal maneuver strategy and avoid collision among satellites. The proposed method is not only able to combine the advantages of two algorithms, but also overcomes the limitations of each algorithm to some extent. More specifically, LPM is used to solve formation reconfiguration with feasible solution in the case of that the distance between any two satellites is far away. The computation time is greatly shortened because there is no collision avoidance constraint. Meanwhile, the CPSO algorithm is applied to enhance solving precision, as well as avoid collision when the satellites stay at a relatively short distance from each other. In order to fully demonstrate the efficiency of proposed method, an array of experiments with comparison to multi-phase LPM are carried out on three satellites. The numerical results verify that proposed method performs well in designing fuel-optimal strategy and ensuring the safety of satellite. It is a feasible and effective algorithm for addressing satellite formation reconfiguration problems.