Incorporating driver expectations into a taxonomy of transfers of control for automated vehicles

Abstract

This study expands existing taxonomies for transfer of control (TOC) with driving automation. A TOC taxonomy is necessary to categorize types of TOC, interpret drivers’ behavior during a transition, and assess the safety implications of each TOC type. However, existing taxonomies do not capture important aspects of the driver’s reaction to the transfer because they often only focus on the initiator and receiver. Thus, we synthesized relevant literature and suggested factors for expanding a TOC taxonomy. The literature synthesis revealed that temporal attributes (e.g., urgency) and predictability of TOC are dimensions that can extend existing taxonomies. Based on the findings, we developed a new TOC taxonomy using four factors (initiator, receiver, system predictability, and driver expectation) with relevant use cases. The new taxonomy distinguishes 16 different types of TOCs and enhances our understanding of TOC types and their implications for safety.