AbstractsTransportation

Authority Transition Interface: A Human Machine Interface for Taking over Control from a Highly Automated Truck:

by C.A.G. Belderbos




Institution: Delft University of Technology
Department:
Year: 2015
Keywords: automation; driver interaction; research through design; authority transitions
Record ID: 1241800
Full text PDF: http://resolver.tudelft.nl/uuid:871045bf-c172-4851-b042-9bc6f82ddc80


Abstract

In recent years, the interest in and development of automated vehicles and its driver supporting systems have significantly grown. Car manufacturers predict to launch the first highly automated cars in the early 2020’s. But before the level of full automation is reached, the control over the vehicle will be shared between the driver and the automation. This sharing of control implies the occurrence of handovers of control and therefore the need for a new human machine interface in these automated vehicles. This thesis is the result of the assignment: “Design an HMI that controls safe (for both driver and traffic), comfortable and efficient authority transitions in a highly automated vehicle using research on new demands for interaction”. The goal was to integrally design an HMI that enables the transitions of control between driver and automated system by supporting the driver in its perception, analysis and decision-making tasks during its changing role in these authority transitions. The conclusions on the analysis are combined and form an HMI framework for ATs. An AT blueprint identifies the general takeover procedure in 7 steps: highly automated driving, take-over or monitoring request, information processing, decision making, task execution, system limit, and lower level of automated driving. The driver reengagement behaviour exists of 4 consecutive steps, with the goal to identify the ‘potential danger in the own lane’, ‘potential danger disregarding location’, ‘destination and potential accident avoidance’, ‘completing mental models’. The information in the road scene and truck layout elements can be prioritized as 1) own lane/road in front far ahead, 2) own lane/road in front close by, 3) left bottom mirror and ICL (dependent on how speed critical the situation is), 4) other mirrors and object in the road scene. The mental models which should be obtained during ATs are ‘current automation state’, ‘current vehicle state’, ‘current driver state’, ‘desired driver state’, ‘close surroundings’, ‘general traffic situation’, ‘landscape and environment’. Furthermore, a list of design guidelines regarding HMIs for ATs is formulated. The Authority Transition HMI divides the tasks and communication between the human operator and automated operator via a structured communication system, categorizing the information and displaying it in three levels in the vehicle cockpit. The concept provides peripheral support in a multimodal and spatial way that adapts to the behaviour of the driver. The interaction of the concept is presented in an interaction framework scenario which defines the AT process by means of a natural driver-automation dialogue. The process is set up as a concise step by step process in which each action by one of the operators should be confirmed by the other. The functionality of the concept is realized though different features which are integrated in the existing vehicle HMI layout. The concept enables the ATs through a collaborative steering wheel, peripheral HUD, haptic seat, speech based sound…