Confidence Horizons: Dynamic Balance of Human and Automation Control Ability in Cooperative Automated Driving

Abstract

This chapter presents the concept of confidence horizon for cooperative vehicles. The confidence horizon is designed to let the automation predict its own and the human’s abilities to control the vehicle in the near future. Based on the pattern approach originating from Alexander et al. [1], the confidence horizon concept is instantiated with a pattern framework. In case of a necessary takeover of the driving task by the human, a mode transition pattern is initiated. In order to determine when the takeover is required, which pattern to start and when to omit the takeover attempt and directly start a minimum risk maneuver, the confidence horizon for both human and co-system is an important parameter. A visual representation of the confidence horizon for the driver in different scenarios prior to a takeover request was explored. Intermediate results of a simulator study are presented, which assess the confidence horizon in automation safety-critical takeover scenarios involving an intersection and a broken-down vehicle on a highway.