Learning to Navigate Under Imperfect Perception: Conformalised Segmentation for Safe Reinforcement Learning

TL;DR

This paper introduces COPPOL, a conformal-based perception method that provides safety guarantees for hazard detection in navigation, significantly improving hazard coverage and reducing violations in real-world benchmarks.

Contribution

The paper presents COPPOL, a novel conformal approach that integrates finite-sample safety guarantees into semantic segmentation for risk-aware reinforcement learning.

Findings

Up to 6x increase in hazard coverage.

Approximately 50% reduction in hazardous violations.

Robust to distributional shifts, maintaining safety and efficiency.

Abstract

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.