Learning to Navigate Under Imperfect Perception: Conformalised Segmentation for Safe Reinforcement Learning
On the map: Deep reinforcement learning methods → Safe and constrained RL
Abstract (arXiv)
Reliable navigation in safety-critical environments requires both accurate hazard perception and principled uncertainty handling to strengthen downstream safety handling. Despite the effectiveness of existing approaches, they assume perfect hazard detection capabilities, while uncertainty-aware perception approaches lack finite-sample guarantees. We present COPPOL, a conformal-driven perception-to-policy learning approach that integrates distribution-free, finite-sample safety guarantees into semantic segmentation, yielding calibrated hazard maps with rigorous bounds for missed detections. These maps induce risk-aware cost fields for downstream RL planning. Across two satellite-derived benchmarks, COPPOL increases hazard coverage (up to 6x) compared to comparative baselines, achieving near-complete detection of unsafe regions while reducing hazardous violations during navigation (up to approx 50%). More importantly, our approach remains robust to distributional shift, preserving both safety and efficiency.
Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Safe and constrained RL, one of 15 directions in Deep reinforcement learning methods. Builds on 19 of its 37 references that are in the map. First posted on arXiv 21 Oct 2025.
Builds on
19 of 37 references.
References · 19
- 2017Constrained Policy Optimizationcited by 983
- 2020A Review of Uncertainty Quantification in Deep Learning: Techniques, Applications and Challengescited by 819
- 2019Conformalized Quantile Regressioncited by 629
- 2017Safe Reinforcement Learning via Shieldingcited by 449
- 2017Safe Model-based Reinforcement Learning with Stability Guaranteescited by 366
- 2020Uncertainty Sets for Image Classifiers using Conformal Predictioncited by 366
- 2018Safe Exploration in Continuous Action Spacescited by 247
- 2015Bayesian SegNet: Model Uncertainty in Deep Convolutional Encoder-Decoder Architectures for Scene Understandingcited by 218
- 2020Bayesian Neural Networks: An Introduction and Surveycited by 131
- 2022Safe Reinforcement Learning via Shielding under Partial Observabilitycited by 56
- 2023Conformal Decision Theory: Safe Autonomous Decisions from Imperfect Predictionscited by 50
- 2024Adaptive Bounding Box Uncertainties via Two-Step Conformal Predictioncited by 23
7 more
- 2023Confident Object Detection via Conformal Prediction and Conformal Risk Control: an Application to Railway Signalingcited by 22
- 2024Conformal Semantic Image Segmentation: Post-hoc Quantification of Predictive Uncertaintycited by 20
- 2023Robust Uncertainty Quantification Using Conformalised Monte Carlo Predictioncited by 10
- 2023State-wise Constrained Policy Optimizationcited by 9
- 2022Hardware faults that matter: Understanding and Estimating the safety impact of hardware faults on object detection DNNscited by 7
- 2022Is it Worth to Reason about Uncertainty in Occupancy Grid Maps during Path Planning?cited by 2
- 2019Learning Autonomous Exploration and Mapping with Semantic Visioncited by 1
Built on by
Not cited yet.
Similar, not cited
Close in wording, no citation either way
- 2026Safety-Constrained Reinforcement Learning with Post-Training Reachability Verification for Robot Navigation
- 2026CoSMO: Conformal Semantic Mapping for Open-World Safe Robot Navigation ↗
- 2025Safety Representations for Safer Policy Learning
- 2025Risk-Guided Diffusion: Toward Deploying Robot Foundation Models in Space, Where Failure Is Not An Option
- 2026Lightweight Safe Reinforcement Learning for End-to-End UAV Navigation
- 2026Safe Reinforcement Learning with Receding-Horizon Safety Critic and Differentiable Action Shields
- 2026Learning Adaptive Safety Margins for Visual Navigation
- 2026ShieldVLA: Feasibility-Aware Safety Alignment for Vision-Language-Action Models