Evaluating Scenario-based Decision-making for Interactive Autonomous Driving Using Rational Criteria: A Survey

TL;DR

The paper tackles the challenge of evaluating DRL-driven decision-making for interactive autonomous driving across highways, ramps, roundabouts, and unsignalized intersections using the DDTUI five-factor framework. It systematically categorizes methods by the number of criteria integrated, from single-factor to five-factor designs, and highlights the growing role of MARL and hybrid control to improve safety, efficiency, and interpretability. Key findings show that very few works address all five factors simultaneously, with interpretability frequently underemphasized, indicating a need for integrated, transparent frameworks. The survey provides a structured map of current approaches, clarifying how scenario-specific demands shape DRL design and offering actionable directions for future research and deployment in real-world traffic systems.

Abstract

Autonomous vehicles (AVs) can significantly promote the advances in road transport mobility in terms of safety, reliability, and decarbonization. However, ensuring safety and efficiency in interactive during within dynamic and diverse environments is still a primary barrier to large-scale AV adoption. In recent years, deep reinforcement learning (DRL) has emerged as an advanced AI-based approach, enabling AVs to learn decision-making strategies adaptively from data and interactions. DRL strategies are better suited than traditional rule-based methods for handling complex, dynamic, and unpredictable driving environments due to their adaptivity. However, varying driving scenarios present distinct challenges, such as avoiding obstacles on highways and reaching specific exits at intersections, requiring different scenario-specific decision-making algorithms. Many DRL algorithms have been proposed in interactive decision-making. However, a rationale review of these DRL algorithms across various scenarios is lacking. Therefore, a comprehensive evaluation is essential to assess these algorithms from multiple perspectives, including those of vehicle users and vehicle manufacturers. This survey reviews the application of DRL algorithms in autonomous driving across typical scenarios, summarizing road features and recent advancements. The scenarios include highways, on-ramp merging, roundabouts, and unsignalized intersections. Furthermore, DRL-based algorithms are evaluated based on five rationale criteria: driving safety, driving efficiency, training efficiency, unselfishness, and interpretability (DDTUI). Each criterion of DDTUI is specifically analyzed in relation to the reviewed algorithms. Finally, the challenges for future DRL-based decision-making algorithms are summarized.

Figures (5)

• Figure 1: Autonomous driving in different scenarios.
• Figure 2: Rear-end accident conditions between ADS and HDV: (a) Rear-end accidents that HDV hit an ADS from behind with a sample of 252; (b) Rear-end accidents that ADS hit an HDV from behind with a sample of 67 [16].
• Figure 3: DRL-based autonomous driving system
• Figure 4: Example scenarios of autonomous driving: (a) highway; (b) on-ramp merging; (c) roundabout with 12 ports (8 entrances: EM1–EM4, EB1–EB4; 4 exits: O1–O4) and a central planted island; (d) unsignalized intersection.
• Figure 5: The importance and necessaries of achieving DDTUI in real-world autonomous driving.