Impulsive guidance of optimal pursuit with conical imaging zone for the evader

This paper addresses the problem of orbital maneuver of the pursuer to position the evader within its imaging zone, which is a spherical cone centered at the pursuer's camera with its centerline parallel to the sunlight. Two pursuit strategies are proposed respectively for the free-floating and maneuvering evader based on geometry analysis of the maneuver trajectory. Specifically, the shortest path from the evader to the conical imaging zone of the pursuer is first calculated and used to analytically design a fixed-time near-fuel-optimal guidance law for the pursuer with a fixed time interval. The magnitude of each impulse is guaranteed to be within the allowable range. Compared with the Sequence Quadratic Programming (SQP) method, the superiority in robustness and computation efficiency of this method over SQP is illustrated. Furthermore, for the maneuvering evader whose maneuver strategy is unknown to the pursuer, the proximal policy optimization (PPO) algorithm in reinforcement learning is applied to train and guide the pursuer to approach the evader intelligently. To enhance the stability and efficiency of the training process, the shortest path from the conical imaging zone to the evader is used to define a unified reward function. Since the reward function synthesizes the relative distance and sunlight angle between the pursuer and evader, reward parameters can be reduced, which is more convenient and efficient for the design and implementation of the reinforcement learning algorithm.