Search papers, labs, and topics across Lattice.
This paper addresses the bottleneck in spacecraft rendezvous and proximity operations (RPO) planning by developing a hierarchical framework that grounds Large Language Model (LLM) reasoning in domain-specific planning modules and a graph of reusable behaviors. The LLM translates natural-language commands into partial mission specifications, which are then resolved by planners and converted into dynamically feasible trajectories via trajectory optimization. Experiments show this architecture achieves 98% exact recovery of partial mission specifications, significantly improving intent recovery over direct LLM generation.
LLMs can now generate dynamically feasible spacecraft rendezvous trajectories by grounding their semantic reasoning in a graph of reusable behaviors and domain-specific planning modules, achieving 98% intent recovery.
Spacecraft rendezvous and proximity operations (RPO) are currently planned through an expertise-intensive process in which engineers translate high-level operational intent into safe, dynamically feasible trajectories, creating a bottleneck to scalable operations. Large language model (LLM)-based agents could offer an intuitive interface for this process, although their outputs are not inherently grounded in orbital dynamics, operational constraints, or the structure of admissible spacecraft maneuvers. To exploit their semantic reasoning while ensuring the generated plan's physical validity, this paper presents a hierarchical framework for spacecraft task-and-motion planning (TAMP) that grounds LLM reasoning in a graph of reusable behaviors and domain-specific planning modules. Within this framework, a pretrained LLM maps a natural-language command to a partial mission specification. The associated planners then resolve unspecified decisions within the admissible operational space. Finally, trajectory optimization converts the completed mission specification into a dynamically feasible trajectory. Numerical experiments demonstrate that this architecture substantially improves intent recovery over direct LLM generation, achieving 98% exact recovery of partial mission specifications across all evaluated splits when backed by frontier LLMs. Additional test-time-compute experiments show that, for a compact 9B model, verifier-guided revision increases exact recovery from 75% to 88%, while broader behavior-plan search independently improves selection among admissible trajectory realizations. Overall, these results establish a scalable and auditable foundation for language-driven agentic planning of spacecraft RPO.
