Quantifying attentional demand of a lane-keeping task as the minimum required information in predictive processing
Year: 2021
Type: conference-paper
Abstract: Automated vehicles (AVs) are expected to improve road safety by reducing the number of collisions and safety critical events [1,2].In the event of a failure of the automated driving system (ADS), or if it reaches the limit of its capability, action is required to preserve the safety of the vehicle occupants and other road users.The Society of Automotive Engineers (SAE) [3] highlights the need for the user to assume control in these situations by performing the driving task or placing the vehicle in a safe state, i.e. one in which the risk of a collision is reduced ('minimal risk condition' (MRC)).Should the driver not be available or deemed unable to achieve this (for example, if they are inattentive or distracted by a non driving-related task (NDRT)), the ADS (operating at level 4 or above) must achieve the MRC.MRC is therefore the subject of intense scrutiny, with many automotive manufacturing companies (OEMs) already proposing solutions, for example, by performing a controlled stop of the car at the side of the road (a 'minimal risk manoeuvre' (MRM)).This is arguably an extension of existing advanced emergency braking systems (AEBS) that sense driver inactivity or their lack of an adequate braking response and intervene.However, one of the key differences in a L4-AV is that drivers are "out of the loop" by design (i.e.not driving and engaged in a NDRT), and the MRM may therefore be entirely unexpected.As such, drivers' responses and behaviour are unpredictable.For example, the ADS will likely prohibit driver intervention during the MRM [3], but it is unclear whether the driver would expect or attempt to intervene and what the consequences might be.These factors are likely to influence the acceptability of proposed solutions.In a medium-fidelity driving simulator, sixteen experienced drivers (10 female, 6 male, 21-65 years old, mean age 31.8,mean driving experience: 12.0yrs with licence, 4.0 hrs/wk driving) undertook three 8-minute drives, each within the same simulated environment created using STISIM Drive software (v3).The driving simulator mimicked an SAE level 4 AV in that it was capable of monitoring the environment and executing both lateral and longitudinal control, thus relieving the driver from primary driving tasks.The MRM was achieved by changing lanes and bringing the vehicle to a stop outside the active traffic lanes (on the "hard shoulder")a commonly proposed 1 DDI2020 -LYON France,
Author: Kujala, Tuomo
Language: English
Subfield: Human Factors and Ergonomics
Field: Social Sciences
Domain: Social Sciences
Open Access status: green
Funders: U.S. Department of Transportation, AAA Foundation for Traffic Safety, European Commission, Transport Accident Commission, Agence Nationale de la Recherche, Tekes, Generalitat Valenciana, State Government of Victoria, Centre National de la Recherche Scientifique, Fondation MAIF, Agencia Estatal de Investigación
OpenAlex ID: W7071024603
