Session Moderators

Moderator(s): 
Chris Schwarz, Director of Engineering & Modeling Research, Driving Safety Research Institute, University of Iowa 

Session Organizers

Organizers:
Chris Schwarz, Driving Safety Research Institute (DSRI)
Kyle Wilson, Seeing Machines
John Gaspar, DSRI
Justin Mason, DSRI
Ensar Becic, National Transportation Safety Board (NTSB)
Rafael Cirino Goncalves, University of Leeds

TRB Sponsoring Committees and Partners

TRB Standing Committee on Impairment in Transportation Committee (ACS50)
TRB Standing Committee on Road User Measurement and Evaluation (ACH50)

Session Description and Agenda

Description

Discussions of driver state monitoring (DSM) systems are typically focused on driver interactions with attention alerts. In contrast, this session explores the premise that DSM systems are capable of much more. Signals from cameras as well as other sensors are used to detect distraction, drowsiness, and intoxication during manual driving. Greater awareness of driver state could be used by ADS to request takeovers, adapt their performance, and even activate automation. Similar ideas are used today in ADAS but higher automation levels have so far focused solely on narrow definitions of driver engagement. Increasingly, DSM systems will provide dual-use functionality for manual and automated driving. This session will invite ideas on DSM from impairment detection to the activation and adaptation of automation and ask what is possible in the larger ADS space. We will identify barriers that touch on technical, policy, liability, and privacy issues; however, overcoming them has the potential to substantially increase safety.

Agenda

The session agenda is a mixture of presentations to frame the issues and small group activities to discuss and debate them. The desired outcome is an identification of research and policy gaps that need to be addressed to advance DSM technology and its applications in Automated Driving Systems (ADS).

We have four highly qualified speakers that will provide background and context to the session: Alexandra Mueller from the Insurance Institute for Highway Safety, Kyle Willson from Seeing Machines, Kathryn Lucaites from NHTSA, and Peter Burns from Transport Canada.

Small groups will form and focus around two types of activities. Before the break, they will be presented with scenarios for using DSM to detect and address some form of impairment during different phases of a ride and in different stages of automation. After the break, they will adopt the points of view of different roles, such as manufacturer, regulator, etc. They will first meet together to discuss the issues and reach consensus on their views. Then they will mix it up with other roles to debate the merits and concerns of widely available DSM technology in ADS. There will be time during the activities to ask questions of and share comments with the larger group.

The session will conclude with brief reporting from the groups and a collection of all written outputs.

Schedule
1:30-1:40 Introduction: Chris Schwarz, Ph.D. (DSRI)
1:40-1:55 Speaker: Alexandra Mueller, Ph.D. (IIHS)
1:55-2:10 Speaker: Kyle Wilson, Ph.D. (Seeing Machines)
2:10-2:40 Small group: scenario analysis, iteration 1
2:40-3:00 Small group: scenario analysis, iteration 2
3:00-3:30 Break
3:30-3:45 Speaker: Kathryn Lucaites, Ph.D. (NHTSA)
3:45-4:00 Speaker: Peter Burns, Ph.D. (Transport Canada)
4:00-4:20 Small group: point-of-view exercise, part 1
4:25-4:45 Small group: point of view exercise, part 2
4:45-5:00 Report out, collect outputs

Session Objectives

The goals of this session are:
-Understand how adjacent applications use DSM
-Consider the extent of ADS responsibility to detect comprehensive driver state
-Explore new ways of applying DSM data
-Expand ideas on inferring driver state (e.g. could use data from wearables, habitual patterns, etc)