ESR 5 – Developing more acceptable, pleasant and transparent AV-kinematic cues for drivers

Popular scientific abstract

Regarding automated vehicles (AVs), technical aspects have been talked a lot, for example, advanced lane-keeping methods. Although it is human drivers that will use AVs, human perspectives are usually neglected, for instance, what kind of driving styles are desired by the user and regarded as comfortable. This is important as desirable and comfortable driving experience will contribute to broad acceptance and usage of AVs.

Even sitting at the driving position, the human driver will lose control over the vehicle. In this case, they cannot operate the car in their personal way by adjusting the preferred speed, acceleration and turns etc. Rather, they have to passively experience the drive operated by the automated driving system. It is unknown how people feel toward this ‘non-self’ driving; can they understand what’s going on with the car? Are they comfortable with the auto-generated driving?

Human-like driving style can be a solution to help the driver to get used to since these driving styles are learnt from humans. Yet, there are a huge number of human driving styles, in which human drivers have different reacting ways toward different situations. This means we will have a huge amount of different combinations of vehicle metrics (velocity, acceleration, lateral offset etc.). What combination is favoured?  As passive drivers, can humans understand the intention of the vehicle when different vehicle metrics applied; for example, the vehicle starts to reduce speed in different distance to a parked car? Which operating style is considered as more desirable and comfortable? Moreover, how can we evaluate comfort as it is a subjective experience?

In summary, my focus is human-like driving; I will investigate if it can bring smoother communication/interaction with humans, as well as contribute to a better experience.

My affiliation

Contact information to supvervisors:

Prof Natasha Merat (Leeds) : N.Merat@its.leeds.ac.uk

Prof Marjan Hagenzieker (TU Delft): m.p.hagenzieker@tudelft.nl

Dr Chongfeng Wei (Northumbria University): chongfeng.wei@northumbria.ac.uk

Background

Manual driving has provided evidence that driving styles influence passengers’ comfort (Bellem, Schönenberg, Krems, & Schrauf, 2016). In automated vehicles, previous active drivers lose control and become passive passengers (Elbanhawi, Simic, & Jazar, 2015). Yet, we do not know how comfortable they feel with regard to automated driving.  Human-like driving styles, imitating humans driving, might be desirable as it might bring familiarity, reliability and avoidance of surprise and make the driving more natural (Ramm, Giacomin, Robertson, & Malizia, 2014). However, it is unknown if the passive driver can understand the intention of the autonomous vehicle as easy as the driver operates by their own even it drives like a human; this communication should contribute to familiarity and less surprise for the driver. Moreover, different driving styles are different combinations of vehicle metrics like speed, acceleration, headway distance, overtaking and violation to traffic law (Hooft van Huysduynen, Terken, Martens, & Eggen, 2015). Humans have different personal driving styles. It is unknown what kind of driving styles that applying different metrics settings are more comfortable. Also, there are no standard evaluations for measuring comfort as it is a subjective feeling.

Aims and objectives

  • To understand whether human-like control models of driving are more comfortable, natural and desirable for users of AVs;
  • To understand how different driver characteristics and different road environments may affect comfort and affect preference of controllers;
  • To investigate how kinematic cues can be used to provide transparent communication of AVs for drivers;
  • To investigate what critical kinematic cues can be used to shape human-like properties of AVs, and contribute to more comfortable experience;
  • To investigate and implement comfort definition, measurement and ways of improvement in AVs;

Research description

  • Become familiar with the research area, by conducting a comprehensive review of published work in this area, preparing an overview report, with an aim to understand the effect of human-like controller on drivers’ levels of comfort;
  • To plan/design suitable experiments for data collection in collaboration with the supervision team;
  • To analyse data from the empirical studies for dissemination at relevant conferences and publication in peer reviewed journals and conference proceedings;

To write up and submit at papers where appropriate to a high-impact journal, such as Human Factors and Ergonomics Society, or Transportation Research Part F, as well as preparing/submitting papers to a highly regarded international conference; 

Results

Expected results:

  • Knowledge on how to make AV kinematics more comfortable (smoother and less jerky), and therefore more acceptable, by improving control and interaction models for AVs;
  • New objective and subjective tools for assessing user comfort and perceived naturalness in AVs;
  • Design guidelines on how to use AV kinematic cues to communicate AV capabilities and limitations transparently.

My publications

To come….

References and links

Bellem, H., Schönenberg, T., Krems, J. F., & Schrauf, M. (2016). Objective metrics of comfort: Developing a driving style for highly automated vehicles. Transportation Research Part F: Traffic Psychology and Behaviour, 41, 45-54. doi:https://doi.org/10.1016/j.trf.2016.05.005

Elbanhawi, M., Simic, M., & Jazar, R. (2015). In the Passenger Seat: Investigating Ride Comfort Measures in Autonomous Cars. IEEE Intelligent Transportation Systems Magazine, 7(3), 4-17. doi:10.1109/MITS.2015.2405571

Hooft van Huysduynen, H., Terken, J., Martens, J.-b., & Eggen, B. (2015). Measuring driving styles: a validation of the multidimensional driving style inventory. Paper presented at the Conference on Automotive User Interfaces and Interactive Vehicular Applications, Nottingham.

Ramm, S., Giacomin, J., Robertson, D., & Malizia, A. (2014). A First Approach to Understanding and Measuring Naturalness in Driver-Car Interaction. Paper presented at the Proceedings of the 6th International Conference on Automotive User Interfaces and Interactive Vehicular Applications, Seattle, WA, USA. https://doi.org/10.1145/2667317.2667416