Human factors of vehicle automation: measuring and understanding driver state during automation

Introduction
In a bid to increase safety, reduce transport-related emissions and increase road capacity, there has been a steady and aggressive rise in the implementation of ‘advanced driver assistance systems’ and automated support systems in vehicles, during the past 5-10 years. The ultimate aim of many vehicle manufacturers and service providers is now to change the state of mobility in the future by providing fully ‘driverless’ or autonomous driving.

As the degree of automation in vehicles increases and the vehicle takes more tasks away from the human driver (keeping the car in the lane and controlling its speed and distance to other cars), the driver starts to engage in other tasks such as catching up on emails or watching TV, etc. The implications of such removal of driver attention from the main driving task are not currently well understood, but could have profound implications on road safety. Investigating whether drivers are still able to comprehend what is happening in the road and driving scene during automation is therefore one of the main concerns of psychologists and human factors specialists studying vehicle automation. It is also not currently clear whether and how drivers can safely and swiftly resume control of the driving task, if they are required to do so. However, research in our laboratories, using driving simulator studies, has shown that drivers are slower at responding to hazards when they are required to resume control from automation and they are also more likely to engage in other, non- driving related tasks during automated driving (Merat et al., 2014; Carsten, et al., 2012).

Project Aim
The aim of this project is to use human psychophysiological measures such as heart monitoring (for assessing alertness, stress and workload) and eye and head tracking (for assessing attention), to establish how vehicle automation affects participants’ engagement in the driving task and to understand what drivers do when automation is engaged. As the automated systems advance in capability, there is less and less call for drivers to be involved in the driving task. However, should the system reach its limitations or the vehicle come to the end of a road where automation is no longer possible, the vehicle needs to ensure that the driver is in a safe and ready state to resume control from the vehicle and re-engage in manual driving. This transition of control back to the driver must be supported by a driver monitoring system which establishes whether the driver is actually alert and capable of resuming control. The proposed research programme will therefore establish what type of monitoring is successful in providing this information and also investigate methods for ensuring the driver is not able to completely disengage from control of the vehicle. Understanding how to keep the driver vigilant, yet not bored of monitoring automation, is also an important consideration of this project.

The project will primarily use the University of Leeds Driving Simulator to investigate the above concepts and co-supervision will be provided by Professor Mike Lenne, Chief Scientist, Human Factors, from Seeing Machines. Members of Seeing Machines have worked with the team at Leeds for a number of years, by providing support for installation and data analysis techniques for one of their key products: the FaceLab eye tracking system. This eye-tracker is currently installed in the driving simulator and used to understand how drivers’ visual attention is affected by different driving scenarios.

Impact of Research
There is clearly high potential impact of this research since the student will be able to work directly with Seeing Machines who is currently working with a number of OEMs with regards to the implementation of such driver monitoring systems in passenger vehicles.

Training
In addition to attending the University of Leeds training programme for PhD students, the student would benefit from training on use of the driving simulator and also use of the eye tracker and BioPac heart monitor. Seeing Machines has also offered internship visits to its premises, subject to funding approval.

Partners and Collaborators
The project is supported in cash and in kind by Seeing Machines.

Entry requirements/necessary background:
The project is suitable for an individual with a background in the behavioural sciences and some knowledge of psychophysiological techniques such as eye trackers will be useful. Programming skills and knowledge of Matlab are desired.

Please visit our LARS scholarship page for more information and further opportunities: https://www.environment.leeds.ac.uk/study/postgraduate-research-degrees/lars-scholarships/