Patricia Delhomme

Fully Automated Driving: Impact of Trust and Practice on Manual Control Recovery.

Objective: An experiment was performed in a driving simulator to investigate the impacts of practice, trust, and interaction on manual control recovery (MCR) when employing fully automated driving (FAD). Background: To increase the use of partially or highly automated driving efficiency and to improve safety, some studies have addressed trust in driving automation and training, but few studies have focused on FAD. FAD is an autonomous system that has full control of a vehicle without any need for intervention by the driver. Method: A total of 69 drivers with a valid license practiced with FAD. They were distributed evenly across two conditions: simple practice and elaborate practice. Results: When examining emergency MCR, a correlation was found between trust and reaction time in the simple practice group (i.e., higher trust meant a longer reaction time), but not in the elaborate practice group. This result indicated that to mitigate the negative impact of overtrust on reaction time, more appropriate practice may be needed. Conclusions: Drivers should be trained in how the automated device works so as to improve MCR performance in case of an emergency. Application: The practice format used in this study could be used for the first interaction with an FAD car when acquiring such a vehicle.

Simulator Training With a Forward Collision Warning System Effects on Driver-System Interactions and Driver Trust

Objective: The study addressed the role of familiarization on a driving simulator with a forward collision warning (FCW) and investigated its impact on driver behavior. Background: Drivers need a good understanding of how an FCW system functions to trust it and use it properly. Theoretical and empirical data suggest that exploring the capacities and limitations of the FCW during the learning period improves operating knowledge and leads to increased driver trust in the system and better driver-system interactions. The authors tested this hypothesis by comparing groups of drivers differing in FCW familiarity. Method: During the familiarization phase, familiarized drivers were trained on the simulator using the FCW, unfamiliarized drivers simply read an FCW manual, and control drivers had no contact with the FCW. During the test, drivers drove the simulator and had to interact with traffic; both familiarized and unfamiliarized drivers used the FCW, whereas controls did not. Results: Simulator familiarization improved driver understanding of FCW operation. Driver-system interactions were more effective: Familiarized drivers had no collisions, longer time headways, and better reactions in most situations. Familiarization increased trust in the FCW but did not raise system acceptance. Conclusion: Familiarization on the simulator had a positive effect on driver-system interactions and on trust in the system. The limitations of the familiarization method are discussed in relation to the driving simulator methodology. Application: Practicing on a driving simulator with driving-assistance systems could facilitate their use during real driving.

Applying an extended theory of planned behavior to predicting violations at automated railroad crossings

Based on an extended Theory of Planned Behavior (TPB, Ajzen, 1985, 1991), we conducted surveys in order to explain and predict violations at a railroad crossing, among pedestrians (n=153) and car drivers (n=151). Measures were made with respect to three chronologically related railroad crossing situations that varied in risk level. The situations were described in scenarios and depicted on photographs. The participants were recruited in the suburbs of Paris, at two automated railroad crossings with four half-barriers. We found that the pedestrians had stronger crossing intentions than did car drivers, especially at the more congested crossing of the two under study. For both categories of road users, intentions and the amount of intention variance explained by the extended TPB factors decreased significantly with risk level. In the most dangerous situations, risk-taking was the most unlikely and the least predictable Self-reported past frequency of crossing against safety warning devices was the main predictor of the intention to commit this violation again, especially among males, followed by the attitude and the injunctive norm in favor the violation. Moreover, car drivers were influenced in their crossing intentions by the descriptive norm. The presence of another vehicle on the tracks when the safety warning devices were activated was perceived not as facilitating, but as an additional risk factor. The discussion addresses the importance of taking into account these determinants of violations in conceiving countermeasures. Our findings could be especially useful for conceiving risk-communication campaigns.

Explaining Senior Drivers’ Road Near Misses Using Both Self-reported and Automatic Collected Data

This paper reports on part of the French MEDOC Project that was supported by Foundation MAIF, Paris, France [grant number RP4-F14140]. The main goal of this project was to understand near miss situations using both objective and subjective methods for data collection: the vehicle dynamics and registered driver’s actions; and the self-reported conditions that led to the occurrence and self-reported road users’ actions to avoid an accident. The collected data allowed for comparing reported ones. The project involved a total of 154 drivers. However, this paper just addresses the group of 27 older drivers. Although older drivers are generally considered safe and cautious drivers, age-related perceptual and cognitive declines might have an impact on self-perception of their own abilities and behavior leading some of them to be under or over-estimators, which will influence their perception of any sudden event. Thus, this study is important to compare their self-reports with the collected data from sensors.