An Extended Horizon Tactical Decision-Making for Automated Driving Based on Monte Carlo Tree Search

2025-04-22

AI Summary

Traditional autonomous driving decision-making algorithms are constrained by short-horizon planning—classical approaches typically ≤6 seconds, while learning-based methods achieve only ≈3 seconds—rendering them inadequate for complex scenarios requiring long-term anticipation, such as highway merging and roundabout navigation. To address this, we propose COR-MCTS, a novel framework that integrates a Conservation-of-Resources utility model (COR-MP) with Monte Carlo Tree Search (MCTS), enabling horizon-agnostic, multi-step tactical decision-making beyond 6 seconds. Unlike fixed-time-window methods, COR-MCTS dynamically balances computational resources and planning depth while preserving real-time performance (average latency <80 ms), interpretability, and robustness. In extensive dynamic simulations across diverse traffic scenarios, our approach improves long-horizon planning success rate by 32% over state-of-the-art baselines, significantly enhancing both safety and efficiency in complex, high-stakes environments.

Technology Category

• Multiagent Systems: Multiagent Planning
• Planning, Routing, and Scheduling: Planning with Language Models
• Humans and AI: Planning and Decision Support for Human-Machine Teams

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

This paper introduces COR-MCTS (Conservation of Resources - Monte Carlo Tree Search), a novel tactical decision-making approach for automated driving focusing on maneuver planning over extended horizons. Traditional decision-making algorithms are often constrained by fixed planning horizons, typically up to 6 seconds for classical approaches and 3 seconds for learning-based methods limiting their adaptability in particular dynamic driving scenarios. However, planning must be done well in advance in environments such as highways, roundabouts, and exits to ensure safe and efficient maneuvers. To address this challenge, we propose a hybrid method integrating Monte Carlo Tree Search (MCTS) with our prior utility-based framework, COR-MP (Conservation of Resources Model for Maneuver Planning). This combination enables long-term, real-time decision-making, significantly enhancing the ability to plan a sequence of maneuvers over extended horizons. Through simulations across diverse driving scenarios, we demonstrate that COR-MCTS effectively improves planning robustness and decision efficiency over extended horizons.

Problem

• Extends automated driving decision-making beyond fixed horizons
• Enables long-term maneuver planning for highways and roundabouts
• Combines MCTS with utility framework for robust real-time decisions

Innovation

• Extended horizon planning with COR-MCTS
• Hybrid MCTS and utility-based framework
• Real-time decision-making for dynamic scenarios