---
title: "Perturbation-Resilient Autonomous Navigation with Distributionally Robust Reinforcement Learning"
authors:
  - "Zhaofan Zhang"
  - "Minghao Yang"
  - "Sihong Xie"
  - "Hui Xiong"
doi: "10.24963/ijcai.2026/847"
url: "https://ijcai.org/proceedings/2026/847"
pdf: "https://ijcai.org/proceedings/2026/0847.pdf"
venue: "Proceedings of the Thirty-Fifth International Joint Conference on Artificial Intelligence"
year: 2026
pages: "7619-7627"
track: "AI and Robotics"
keywords:
  - "AIR: Robot control, planning, and execution with guarantees"
  - "Robot control, planning, and execution with guarantees: Safe and robust control under uncertainty"
  - "Safety, trustworthiness, generalizability, and evaluation: Theoretical and algorithmic guarantees on safety, robustness, and out-of-distribution generalization"
  - "AIR: Generative AI, robotic foundation models, and reinforcement learning"
---

# Perturbation-Resilient Autonomous Navigation with Distributionally Robust Reinforcement Learning

Zhaofan Zhang, Minghao Yang, Sihong Xie, Hui Xiong

*Proceedings of the Thirty-Fifth International Joint Conference on Artificial Intelligence*, pages 7619-7627. DOI: [10.24963/ijcai.2026/847](https://doi.org/10.24963/ijcai.2026/847)

## Abstract

The robustness of autonomous vehicles such as drones and Unmanned Surface Vehicles (USV) is crucial when facing unknown and complex marine environments, especially when heteroscedastic observational noise poses significant challenges to sensor-based navigation tasks. Recently, Distributional Reinforcement Learning (DistRL) has shown promising results in some challenging autonomous navigation tasks without prior environmental information. However, these methods overlook situations where noise patterns vary across different environmental conditions, hindering safe navigation and disrupting the learning of value functions. To address the problem, we propose DRIQN to integrate Distributionally
Robust Optimization (DRO) with implicit quantile networks to optimize worst-case performance under natural environmental conditions. Leveraging explicit subgroup modeling in the replay buffer, DRIQN incorporates heterogeneous noise sources and target robustness-critical scenarios. Experimental results based on the risk-sensitive environment demonstrate that DRIQN significantly outperforms state-of-the-art meth-
ods, achieving +13.51% success rate, -12.28% collision rate and +35.46% for time saving, +27.99% for energy saving, compared with the runner-up.