Enhancing Social Decision-Making of Autonomous Vehicles: A Mixed-Strategy Game Approach With Interaction Orientation Identification

Jiaqi Liu,Xiao Qi,Peng Hang,Jian Sun

2024-04-04

Abstract:The integration of Autonomous Vehicles (AVs) into existing human-driven traffic systems poses considerable challenges, especially within environments where human and machine interactions are frequent and complex, such as at unsignalized intersections. To deal with these challenges, we introduce a novel framework predicated on dynamic and socially-aware decision-making game theory to augment the social decision-making prowess of AVs in mixed driving environments. This comprehensive framework is delineated into three primary modules: Interaction Orientation Identification, Mixed-Strategy Game Modeling, and Expert Mode Learning. We introduce 'Interaction Orientation' as a metric to evaluate the social decision-making tendencies of various agents, incorporating both environmental factors and trajectory characteristics. The mixed-strategy game model developed as part of this framework considers the evolution of future traffic scenarios and includes a utility function that balances safety, operational efficiency, and the unpredictability of environmental conditions. To adapt to real-world driving complexities, our framework utilizes a dynamic optimization framework for assimilating and learning from expert human driving strategies. These strategies are compiled into a comprehensive strategy library, serving as a reference for future decision-making processes. The proposed approach is validated through extensive driving datasets and human-in-loop driving experiments, and the results demonstrate marked enhancements in decision timing and precision.

Robotics

What problem does this paper attempt to address?

The paper addresses the issue of how autonomous vehicles (AVs) make social decisions in mixed traffic environments with human-driven vehicles. The study proposes a dynamic social decision-making game framework consisting of three modules: interaction tendency identification, mixed strategy game modeling, and expert mode learning. The framework evaluates interaction tendencies by quantifying environmental factors and trajectory features, establishes a game model considering future state evolution and balancing safety and efficiency through a utility function, and learns and adapts expert driving strategies from real-world data to enhance the social interaction capability of AVs. Experiments show that this method outperforms baseline methods in terms of decision timing and accuracy.