Explainable Underlying Mechanism of the Driver's Decision‐Making Process

ABSTRACT As autonomous vehicles increasingly coexist with human‐driven vehicles, the ability to interact intelligently with other road users becomes essential. Existing approaches to modelling such interaction have largely focused on reproducing the final driving behaviours, which limits their ability to explain why specific decisions arise. This paper proposes an interactive decision‐making mechanism for driving agents based on descriptive theories of human reasoning and negotiation. The mechanism captures drivers as rational agents who form beliefs, activate desires and negotiate conflicting plans through vehicle behaviour interpreted as locution. The proposed framework is applied to lane‐merging scenarios, and validation using real traffic data demonstrates that the resulting decision sequences capture key human‐like interaction patterns across diverse situations. This mechanism provides interpretable reasoning traces, enabling explanation of the internal processes that lead to each decision. This process‐oriented perspective offers a principled foundation for understanding interactive driving behaviour and supports the development of autonomous vehicles capable of transparent, human‐like decision‐making.