Home // AISyS 2025, The Second International Conference on AI-based Systems and Services // View article

Risk-Aware HTN Planning Domain Models for Autonomous Vehicles and Satellites

Authors:
Ebaa Alnazer
Ilche Georgievski
Marco Aiello

Keywords: autonomous vehicles; satellites; HTN planning; knowledge engineering; domain models; risk; uncertainty

Abstract:
The real world is characterised by uncertainty and risks. When modelling it as a domain for planning systems, this translates into action outcomes that can not be fully anticipated. In such environments, automating planning requires not only sophisticated algorithms but also domain models that adequately capture such complexity and unpredictability. However, existing AI planning domains often oversimplify these complexities, either because they are designed as benchmarks to evaluate planners or because they were created to test specific methods, frequently at the cost of broader realism. Autonomous vehicles and satellites are representative application examples that pose common planning challenges in dynamic, uncertain environments. Taking these, current domain models frequently omit critical features, such as uncertainty, risk, and the wide range of choices available to agents in achieving their objectives. Here, we contribute towards bringing these domains closer to reality by following a systematic approach to knowledge engineering and domain modelling that better captures these neglected aspects. Our models are implemented within the risk-aware Hierarchical Task Network (HTN) planning framework, which aligns with human-like reasoning and accommodates uncertainty and risk. By enhancing the realism of these two domains, our work increases their relevance for practical applications. Also, this work aims to drive the development of more capable AI planners and encourage the creation of more realistic domain models.

Pages: 36 to 45

Copyright: Copyright (c) IARIA, 2025

Publication date: September 28, 2025

Published in: conference

ISBN: 978-1-68558-303-3

Location: Lisbon, Portugal

Dates: from September 28, 2025 to October 2, 2025