Figure 1: Physiological measures in HAV simulator
Fully automated vehicles (L4 AVs), still in development, will be a near future reality. Many experts expect this type of vehicles to be affordable and commonplace in the 2040s (Böhm, Kocur et al., 2017), while the 69% respondents of another study estimated that fully automated driving will reach a 50% market share between now and 2050 (Kyriakidis, 2015).
Traditionally, in the automotive sector, technology-led innovation has overshadowed human-centered design, and vehicles have been historically designed for male first. This is a key reason for women’s lower willingness to use automated cars (Pflugfelder, 2018). One example are Crash test dummies used in auto safety testing, in which average male bodies or a scaled version of male bodies, are used for females crash tests, which is not representative of female bodies.
One of DIAMOND project’s research studies, leading with women fairness in transport and mobility in general, focus on autonomous vehicles. This study, led by the Instituto de Biomecánica de Valencia (IBV), puts the focus on user-centered testing and design for a fair woman design of autonomous vehicles, with the goal to improve the emotional experience and increase women’s level of acceptance and trust towards autonomous driving technology.
DIAMOND study on autonomous vehicles aims to improve women’s level of acceptance towards autonomous driving technology. While men are more likely to associate positive emotions towards automated cars, women are prone to associate them with anxious emotions
Affective reactions have demonstrated their relevance in explaining the sex differences in the willingness to use automated cars, measured by a questionnaire research (Hohenberger et al, 2016). Men are more likely to associate positive emotions, pleasure, towards automated cars and women are more likely to associate anxiety. Consequently, women’s willingness to use automated cars is lower than men. (Hohenberger et al, 2016)
The main goals of DIAMOND experiments were the identification of gender-driven stressors related to autonomous driving emotional experience, obtaining objective emotional indicators during fully automated driving.
For this purpose, experiments took place in IBV’s Human Autonomous Vehicle (HAV), a dynamic driving simulator. The HAV allows to emulate a vehicle with different degrees of autonomy, its dynamical behavior and to monitor and detect the emotions of the occupants.
The experimentation included a total of 40 drivers between 25-55 years old, half women and half men, with and without care mobility responsibilities linked to having or not children under 12. We wanted to make sure that a group of participants was performing care mobility travels as this type of mobility has a high bias (Criado-Perez, 2019), being mostly performed by women. In each group there was a homogeneous age distribution. The aim of the work was to elicit and measure the emotional reactions of participants with AVs driving, while exploring the influence of gender, age and having or not children.
Six scenarios were designed and implemented in the HAV, with the help of a focus group, considering not only different environmental, traffic and driving conditions but also gender and other relevant intersectional variables in transport.
The first scenario simulated favorable and unfavorable weather conditions and wanted to measure the trust with the autonomous vehicle under these different situations. The second and third scenarios proposed two different driving modes: (1) sport, including an abrupt overtaking, a strong braking and high speeds and accelerations and (2) comfort, with smooth speeds and accelerations. In the fourth scenario, the participants had to perform a non-driving related task involving the use of a tablet. Finaly, the fifth scenario implied a sudden failure of the system and the last one incorporated a dangerous driving situation, in which a car skipped a stop signal, stirring up emotions while the safety perception was being tested.
Drivers’ emotions were estimated by a bidimensional representation of Arousal, the level of activation of the participant, and ranges from calm (or low) to excited (or high), and Valence, the level of pleasantness defined along a continuum from negative to positive, (Russell, 1980) through the recording by sensors of bio-signal data (Figure 2) including facial Electro Myography sensors (EMG), electrocardiogram measures (ECG) and Electro Dermal Activity sensors (EDA).
Figure 2: (left) facial Electro Myography sensors, EMG (center) sensor for obtaining the electrocardiogram, ECG (right) Electro Dermal Activity sensors, EDA.
Results demonstrated that the scenario design elicited differences in the emotional state between genders, in particular in the arousal level. This can be seen specially in the scenarios involving different driving modes and performing non-driving related activities.
Different driving scenarios triggered different emotional responses in drivers between genders
The driving mode (sport vs comfort) impacted the physiological emotional state of both genders. The sport driving mode provoked a higher arousal on participants than the comfort driving mode. However, in men this impact is higher. Men, in the comfort scenario, are significantly less activated (lower arousal) than in the rest of scenarios compared to women. Thus, we find a clear “relaxing” effect of this scenario for men, while we do not find it for women, for whom the comfort scenario has the same arousal level as the others scenarios. Therefore, the driving mode impacts the physiological emotional state of both genders in an expected way although the impact is higher in men.
The scenario involving the performance of non-driving related task provoked also emotional differences. Men, when performing simultaneous tasks, present the highest activation level compared to the other scenarios. In this driving scenario, the activation level for men is higher than the measured for women. This fact could be attributed to a higher cognitive effort of men while performing a simultaneous task.
This experiment validated the hypothesis that it is possible to find differences due to gender in the emotional state of the passengers, considering the physiological signals discussed above. The use of emotional information, combined with other variables related with acceptance such as perceived safety, social impact and personal convenience, is a valuable input for enhancing women’s autonomous vehicles acceptance and boosting a fair design of automated driving.
Instituto de Biomecánica de Valencia
The Instituto de Biomecánica de València (IBV) is a technological center that studies the human behavior and the physical, cognitive and emotional user interaction with products, environments and services, with the aim to improve competitiveness among the business sector. IBV contributes to the project with advanced methodologies for evaluating user interaction and generating design criteria and requirements based on a human centered approach, to enhance the in-vehicle woman experience and to increase acceptance in the case of autonomous vehicles.
References
Böhm, P., Kocur, M., Firat, M., & Isemann, D. (2017, September). Which Factors Influence Attitudes Towards Using Autonomous Vehicles?. In Proceedings of the 9th International Conference on Automotive User Interfaces and Interactive Vehicular Applications Adjunct (pp. 141-145).
Criado-Perez, C. (2019). Invisible women: Exposing data bias in a world designed for men. Great Britain: Vintage Publishing. Retrieved from. https://www.overdrive.com/search?q=91115E7D-9D96-434F-9996-E60D55BDF481
Hohenberger, C., Spörrle, M., & Welpe, I. M. (2016). How and why do men and women differ in their willingness to use automated cars? The influence of emotions across different age groups. Transportation Research Part A: Policy and Practice, 94, 374-385.
Kyriakidis, M., Happee, R., & de Winter, J. C. (2015). Public opinion on automated driving: Results of an international questionnaire among 5000 respondents. Transportation research part F: traffic psychology and behaviour, 32, 127-140.
Pflugfelder, E. H. (2018). Autonomous Vehicles and Gender: A Commentary. Transfers, 8(1), 104-111.
Russell, J. A. (1980). A circumplex model of affect. Journal of personality and social psychology, 39(6), 1161.
