• Event-triggered distributed control in autonomous spacecraft swarms. • Behavior-based local control for target autonomous search, chase, safe enclosure. • Distributed local sensing detect target motion and geometry collectively. • Robust reconstruction handles noisy, partial landmark observations. • Simulations demonstrate feasibility and scalability of swarm control. This paper presents an event-triggered behavior-inspired decentralized control framework for a spacecraft swarm tasked with capturing an unknown, uncooperative target. The proposed approach integrates local swarm behaviors, event-triggered communication and control, and distributed target perception within a unified architecture. Three local behaviors are considered: inter-agent collision avoidance, target buffering, and sensor-pointing alignment. To reduce communication burden, each agent updates and broadcasts its state only when predefined local triggering conditions are satisfied. Meanwhile, onboard sensing and neighbor-to-neighbor information exchange are used to estimate the target position, motion, and geometry without prior knowledge of the target shape. A graduated non-convexity-based perception module is adopted to improve the robustness of landmark-based target reconstruction under partial observability and noisy measurements. Numerical simulations for tumbling ellipsoidal targets demonstrate that the proposed method enables autonomous search, chase, and pre-capture enclosure while maintaining safe inter-agent spacing and reducing communication compared with periodic update strategies.
Zhang et al. (Fri,) studied this question.