Distribution-Agnostic Robust Trajectory Optimization via Chance-Constrained Reinforcement Learning

2026-06-11 · Source: Takara TLDR - Daily AI Papers · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Robotics & Autonomous Systems · Depth: Expert, medium

Summary

A new distribution-agnostic robust trajectory-optimization framework is presented, employing chance-constrained reinforcement learning to manage uncertainty from initial conditions and process noise, requiring only sampleability. The system first computes a deterministic nominal trajectory offline. Reinforcement learning then robustifies this baseline through a structured affine closed-loop correction law, incorporating a feedforward control adjustment and time-varying feedback gains. Probabilistic feasibility is empirically enforced using rollout-based upper-tail quantiles, while terminal dispersion is regulated via covariance-feasibility penalties. The framework was assessed on a three-dimensional multi-impulse Earth-Mars transfer and a stochastic atmospheric pinpoint rocket landing. Results indicate competitive upper-tail fuel cost, preserved probabilistic feasibility, and the robustification scaffold's portability across heterogeneous spacecraft trajectory planning problems without core stochastic-control redesign.

Key takeaway

For aerospace engineers designing robust trajectories for complex spacecraft missions, this framework offers a compelling solution. It effectively manages diverse uncertainties, from initial conditions to process noise, while maintaining probabilistic feasibility and competitive fuel costs. You should consider integrating this distribution-agnostic, chance-constrained reinforcement learning approach to enhance mission reliability and reduce the need for extensive redesign across different trajectory planning problems.

Key insights

A chance-constrained reinforcement learning framework robustifies spacecraft trajectory optimization against diverse uncertainties using an affine closed-loop correction.

Principles

• Uncertainty requires only sampleability.
• Robustify nominal trajectories with RL.
• Enforce feasibility via upper-tail quantiles.

Method

Compute a nominal trajectory offline. Apply reinforcement learning to derive an affine closed-loop correction law. Enforce probabilistic feasibility using rollout-based upper-tail quantiles and regulate terminal dispersion via covariance-feasibility penalties.

In practice

• Apply to multi-impulse Earth-Mars transfers.
• Adapt for atmospheric rocket landings.
• Re-use core structure for diverse spacecraft.

Topics

• Trajectory Optimization
• Reinforcement Learning
• Chance-Constrained Control
• Robust Control
• Spacecraft Guidance
• Stochastic Control

Best for: Research Scientist, AI Scientist, Robotics Engineer

Related on AIssential

• Robust Exploratory Stopping under Ambiguity in Reinforcement Learning — A robust RL framework for optimal stopping under ambiguity balances exploration and exploitation using $g$-expectation and BSDEs.
• Rule-based High-Level Coaching for Goal-Conditioned Reinforcement Learning in Search-and-Rescue UAV Missions Under Limited-Simulation Training — A hybrid rule-based and RL framework enhances UAV mission safety and efficiency in limited-simulation SAR.
• Using Reward Uncertainty to Induce Diverse Behaviour in Reinforcement Learning — Diversity in RL naturally emerges from treating reward as a distribution, not a scalar.
• Optimistic Policy Learning under Pessimistic Adversaries with Regret and Violation Guarantees — RHC-UCRL enables safe, optimal reinforcement learning against explicit adversarial policies with theoretical guarantees.
• WOMBET: World Model-based Experience Transfer for Robust and Sample-efficient Reinforcement Learning — WOMBET unifies uncertainty-aware model-based data generation with adaptive online fine-tuning for efficient RL experience transfer.
• Transit Network Design with Two-Level Demand Uncertainties: A Machine Learning and Contextual Stochastic Optimization Framework — Integrating machine learning with contextual stochastic optimization improves transit network design by modeling two levels of demand uncertainty.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Takara TLDR - Daily AI Papers.