Multiobjective optimization for collision avoidance maneuver using a genetic algorithm

In this paper, we analyze an evolutionary multiobjective heuristic algorithm for collision avoidance maneuver optimization when multiple objects in space threaten a satellite. Population-based evolutionary multiobjective heuristic algorithms can stably find optimal or suboptimal solutions on account of their robustness and flexibility. We present a collision avoidance maneuver optimization planning method using the nondominated sorting genetic algorithm II. Our proposed method simultaneously optimizes two goals: the minimization of DeltaV consumption and maximization of the satellite maneuver cycle length. Through an experimental evaluation, we demonstrate the efficient performance of our method in terms of DeltaV consumption and maneuver cycle length under constraints, including allowable DeltaV and burn start time, with consideration of a safety threshold between the user satellite and multiple threatening objects. We perform the simulation once per each burn strategy, i.e. in an in-track direction with a single burn; in three radial, in-track, cross-track direction with a single burn; and three radial, in-track, cross-track directions with two burns. We obtain various solutions, all of which satisfy the safety threshold. For the in-track direction with the single burn strategy, the solution showed merit in the minimization of DeltaV consumption. On the other hand, the results of three radial, in-track, cross-track directions with two burns showed good performance in maximization of the maneuver cycle. We therefore contend that efficient collision avoidance maneuver optimization planning can be achieved by using a multiobjective heuristic algorithm.