Research Analyzer
← Back ICRA 2026

Monitoring Autonomous Persistent Surveillance Missions Using Invariance

Vladislav Nenchev, Prodromos Sotiriadis

PDF

AI summary

Key figure (auto-extracted from paper)
A compositional invariant-based monitor enables scalable, real-time verification of persistent surveillance missions with black-box autonomy, avoiding the curse of dimensionality while guaranteeing soundness and completeness.
Runtime monitoring Persistent surveillance Invariant sets Hybrid systems Black-box autonomy Compositional verification

Problem

Autonomous robots performing persistent surveillance often rely on black-box control stacks, making it difficult to detect stalls, faults, or adversarial perturbations before they cause operational failure. Existing monitoring methods struggle to scale to large environments or require white-box access and heavy online computation.

Approach

The authors model the robot and region uncertainties as a state-dependent hybrid system and compute per-part Robust Controlled Invariant Sets (RCIS) offline. These local invariants are intersected online to form a scalable monitor that checks state membership in constant time per step.

Key results

  • Formalized persistent surveillance with black-box control as a hybrid LPV system
  • Proved soundness and completeness of a compositional RCIS construction that scales linearly with environment partitions
  • Developed an online monitor with polyhedral membership checks requiring only ~0.03 ms per step
  • Validated the approach on a real differential-drive robot persistently surveying a labyrinth environment

Why it matters

Enables reliable, low-latency safety monitoring for autonomous robots operating in complex, partitioned environments without requiring access to internal control algorithms.

Abstract

This paper studies runtime monitoring for per- sistent surveillance by autonomous robots when the autonomy stack is a black box. The environment is partitioned into finitely many parts, each carrying an uncertainty state that decreases when observed and increases otherwise. We model the closed loop as a state-dependent hybrid system with linear parameter varying dynamics and design a monitor based on an invariant computed offline. As this invariant is typically hard to obtain for large to-be-surveyed spaces, we propose a compositional moni- tor obtained by decentralized computation of low-dimensional invariant sets for each uncertainty region, and checking their conjunction online. Under common independence assumptions, the compositional monitor is sound and complete with respect to the full-system invariant. The approach is applied in a case study with a real robot persistently monitoring a labyrinth, emphasizing its applicability in practice.

Index terms

Hybrid Logical/Dynamical Planning and Verification Failure Detection and Recovery Environment Monitoring and Management

Related papers