Context-aware, Ante-hoc Explanations of Driving Behaviour
Dominik Grundt, Ishan Saxena, Malte Petersen, Bernd Westphal, Eike Möhlmann
TL;DR
The paper tackles the challenge of making AI-driven autonomous driving decisions interpretable and trustworthy by embedding Explainability Engineering (EE) within a runtime framework. It leverages Traffic Sequence Charts (TSC) to formally specify explanation contexts and (un)expectable driving manoeuvres, and employs dedicated runtime monitoring to trigger ante-hoc explanations as the AV operates. Grounded in the MAB-EX framework, the approach operationalises the Monitoring, Analyse, Build, and EXplain phases, enabling context-aware communication tailored to stakeholders. A simulated overtaking scenario in CARLA demonstrates that context-aligned explanations can be presented before manoeuvres, supporting transparency, traceability, and regulatory alignment while highlighting paths for future work in uncertainty handling and extension with explanation history.
Abstract
Autonomous vehicles (AVs) must be both safe and trustworthy to gain social acceptance and become a viable option for everyday public transportation. Explanations about the system behaviour can increase safety and trust in AVs. Unfortunately, explaining the system behaviour of AI-based driving functions is particularly challenging, as decision-making processes are often opaque. The field of Explainability Engineering tackles this challenge by developing explanation models at design time. These models are designed from system design artefacts and stakeholder needs to develop correct and good explanations. To support this field, we propose an approach that enables context-aware, ante-hoc explanations of (un)expectable driving manoeuvres at runtime. The visual yet formal language Traffic Sequence Charts is used to formalise explanation contexts, as well as corresponding (un)expectable driving manoeuvres. A dedicated runtime monitoring enables context-recognition and ante-hoc presentation of explanations at runtime. In combination, we aim to support the bridging of correct and good explanations. Our method is demonstrated in a simulated overtaking.