Distribution-Agnostic Robust Trajectory Optimization via Chance-Constrained Reinforcement Learning
Abstract
This paper presents a distribution-agnostic robust trajectory-optimization framework based on chance-constrained reinforcement learning. The uncertainty is represented here through initial conditions and process noise, with the only requirement being that it can be sampled. A deterministic nominal trajectory is first computed offline, and reinforcement learning is then used only to robustify that baseline through a structured affine closed-loop correction law comprising a feedforward control adjustment and time-varying feedback gains. Probabilistic feasibility is enforced empirically through rollout-based upper-tail quantiles, while terminal dispersion is regulated through covariance-feasibility penalties. The framework is assessed on two materially different trajectory design problems. The flagship case study is a three-dimensional multi-impulse Earth-Mars transfer, where the learned policy is benchmarked against a recent robust trajectory-optimization reference under Gaussian uncertainty and then evaluated under bounded uniform uncertainty and under process disturbances not seen during training. The second case study is a stochastic atmospheric pinpoint rocket landing problem, used to assess portability to a short-horizon continuous-thrust setting with drag, mass depletion, and glide-slope constraints. The results show that the proposed framework can remain competitive in upper-tail fuel cost while preserving probabilistic feasibility, and that the same robustification scaffold can be carried across heterogeneous spacecraft trajectory planning problems without redesign of its core stochastic-control structure.
Versions
Related
- Efficient Joint Trajectory Optimization and Resource Allocation for Secure UAV Networks: A Deep Reinforcement Learning Approach
- Multi-Objective Reinforcement Learning Using Routed Ensembles and Trajectory Attribution
- Target-Constrained Pooling for Trajectory Extrapolation
- Back to the Definition: Estimating Step-Level Advantages via Trajectory Graphs for Agentic Reinforcement Learning
- Constrained Reinforcement Learning-Based Framework for DC Microgrid: Current Sharing and Voltage Regulation
BibTeX
@misc{chaudhary2026distributionagnostic,
title = {{Distribution-Agnostic Robust Trajectory Optimization via Chance-Constrained Reinforcement Learning}},
author = {Yashdeep Chaudhary and Roberto Armellin and Harry Holt and Marco Sagliano},
howpublished = {arXiv},
year = {2026},
month = jun,
doi = {10.48550/arxiv.2606.13605},
eprint = {2606.13605},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2606.13605},
}
RIS
TY - UNPB
TI - Distribution-Agnostic Robust Trajectory Optimization via Chance-Constrained Reinforcement Learning
AU - Yashdeep Chaudhary
AU - Roberto Armellin
AU - Harry Holt
AU - Marco Sagliano
PY - 2026
DA - 2026/06/11
PB - arXiv
DO - 10.48550/arxiv.2606.13605
UR - https://arxiv.org/abs/2606.13605
AB - This paper presents a distribution-agnostic robust trajectory-optimization framework based on chance-constrained reinforcement learning. The uncertainty is represented here through initial conditions and process noise, with the only requirement being that it can be sampled. A deterministic nominal trajectory is first computed offline, and reinforcement learning is then used only to robustify that baseline through a structured affine closed-loop correction law comprising a feedforward control adjustment and time-varying feedback gains. Probabilistic feasibility is enforced empirically through rollout-based upper-tail quantiles, while terminal dispersion is regulated through covariance-feasibility penalties. The framework is assessed on two materially different trajectory design problems. The flagship case study is a three-dimensional multi-impulse Earth-Mars transfer, where the learned policy is benchmarked against a recent robust trajectory-optimization reference under Gaussian uncertainty and then evaluated under bounded uniform uncertainty and under process disturbances not seen during training. The second case study is a stochastic atmospheric pinpoint rocket landing problem, used to assess portability to a short-horizon continuous-thrust setting with drag, mass depletion, and glide-slope constraints. The results show that the proposed framework can remain competitive in upper-tail fuel cost while preserving probabilistic feasibility, and that the same robustification scaffold can be carried across heterogeneous spacecraft trajectory planning problems without redesign of its core stochastic-control structure.
ER -