A dimension reduction guidance method for orbital pursuit-evasion games

Abstract

Improving on-orbit service techniques with non-cooperative spacecraft is attracting more and more scholars' attention. The free time orbital pursuit-evasion problem is a typical problem in this field. Generally, solving the orbital pursuit-evasion game results in solving a 24-dimension two points boundary value problem which is really complicated with such high dimension and the numerical methods used to obtain the solution is sensitive to the initial estimate of the costates. To overcome the weaknesses, this paper presents a dimension reduction guidance method to transform the free-time orbital pursuit-evasion game into a 3-dimension parameters optimization problem based on the costate normalization technique where each of the parameters needs to be optimized has a strict boundary. Comparing to the previous methods, the decrease of the dimension and the boundary of the parameters reduce the searching space together. The simulation results and the statistical analysis demonstrate that this method is efficient and it improves the computation consumption 67.7% comparing with the classical genetic algorithm.