Runtime Monitoring of Perception-Based Autonomous Systems via Embedding Temporal Logic

Parv Kapoor, Abigail Hammer, Ashish Kapoor, Karen Leung, Eunsuk Kang

arXiv:2605.12651 · 2026-05-14 공개 · arXiv · PDF

runtime-monitoring autonomous-systems conformal-calibration temporal-operators embedding-spaces semantic-regions perceptual-behaviors manipulation-environments

Abstract

Runtime monitoring of autonomous systems traditionally relies on mapping continuous sensor observations to discrete logical propositions defined over low-dimensional state variables. This abstraction breaks down in perception-driven settings, where such mappings require additional learned modules that are often computationally expensive, brittle, and semantically misaligned. In this work, we propose Embedding Temporal Logic (ETL), a temporal logic that performs monitoring directly in learned embedding spaces. ETL defines predicates through distances between observed embeddings and target embeddings derived from reference observations. This formulation allows specifications to capture high-level perceptual concepts, such as similarity to visual goals or avoidance of semantic regions, that are difficult or impossible to express using traditional predicates. By composing these predicates with temporal operators, ETL naturally expresses temporally extended and sequential perceptual behaviors. We introduce ETL monitors for evaluating specifications over bounded embedding traces, along with a conformal calibration procedure that provides reliable and safety-oriented predicate evaluation. We evaluate our approach across multiple manipulation environments to show that ETL achieves strong empirical agreement with ground-truth semantics, including accurate monitoring of temporally composed behaviors.

한국어 요약

📋 한 줄 요약

**[자율 시스템 / 형식 검증]** 학습된 임베딩 공간에서 직접 시간 논리식을 평가하는 새로운 런타임 모니터링 프레임워크 Embedding Temporal Logic(ETL)을 제안.

🎯 핵심 기여도

💡 핵심 아이디어

인지 기반 자율 시스템에서는 저차원 명제로의 변환이 부서지므로, 형식 명세를 **임베딩 공간 자체**에서 정의하는 편이 자연스럽고 강건하다. 거리 기반 술어가 비전·언어 모델의 고수준 의미를 그대로 활용할 수 있게 해 준다.

🔬 기술적 접근법

📊 주요 결과

💭 의의 및 한계

**의의**: 인지 기반 자율 시스템(로봇·자율주행)에서 명세-실행 간 의미 갭을 좁히는 실용적 형식 모니터링 도구. **한계**: 임베딩 공간 자체가 갖는 학습된 편향에 술어 의미가 의존하므로, 임베더 교체 시 명세 일관성 유지 메커니즘은 추가 연구 필요.

🚀 실용적 활용