A tool for training hazard perception and for assessing driving behaviors in adolescents and inexperienced drivers: A simulation study

All over the world, young and inexperienced road users are overrepresented among the victims of crashes. They are reported to have a worse hazard perception, being at the same time more prone to risk-taking behaviors. Moreover, various patterns of personality traits seem to affect their on-road behaviors. During the years, lots of efforts have been devoted to the identification of strategies to cope with such a phenomenon, usually employing psychosocial interventions and frontal lessons. Meanwhile, the progress in technology led to the development of driving simulators, allowing to test driving behaviors in a safe environment. However, to date no specific training or assessment driving simulator protocols have yet been developed. The present Ph.D. project aimed at developing and testing a hazard perception training and a driving profile assessment protocol for inexperienced (young and adolescents) road users by mean of a specific driving simulator, psychophysiological indices, and psychological measures. The development of the training and of the assessment protocol followed several steps, sometimes running parallel to each other. The features and issues of the inexperienced road users are discussed in Chapter 1, where a distinction between experience and exposure is made and their links with onroad behaviors are explored. In the second part of Chapter 1, the pros and cons of driving simulators are described; Moreover, it is introduced the Honda Riding Trainer - HRT simulator (employed in this Ph.D. project), followed by an overview of the previous studies that have previously used this simulator. In Chapter 2 the most important factors influencing inexperienced road users’ behaviors (hazard perception, attention, and personality traits) are introduced and discussed. The research work can be divided into two main branches: The first (Chapter 3) includes three studies. Specifically, the first two constituted the two parts of a single research project aimed at assessing changes in the skin conductance (SC) considered as a somatic marker of hazard perception in novice drivers during a two and a three-session training on the HRT, respectively. In the case of a two-session training, the main results show that the reduction of the crashes while driving the simulator (when participants faced again the same courses during the second session) is paralleled by a significant anticipation in the onset of skin conductance responses (SCRs) in the hazardous scenes included in the courses, testifying that participants learned to anticipate the hazards that they had already experienced. In the subsequent study, three sessions were employed (the first two with the same courses and the last with different ones) and two groups compared: The training group drove the simulator whereas the control group just had to watch the video-clips of the same courses and detect hazards. Here, the reduction in risky driving behaviors and the increase in the safe ones in the training group were paralleled by significant changes in the SC in terms of percentages of SCRs, but no anticipation of the SCR onset was present in the third session. The last study included in Chapter 3 aimed at assessing the event-related potentials (ERPs) of attentional monitoring and shifting in a one-session training on the HRT, during which participants were administered with a multi-feature oddball paradigm. The deviant stimuli in the oddball were divided into three categories: Deviant pure tones, human sounds and traffic-related sounds. The main results showed that, at the end of the training, the participants avoided a higher percentage of accidents and had enough cognitive resources available to monitor the environment and to shift attention toward salient stimuli (traffic-related sounds), as attested by the increase in the amplitude of the P3a. The second branch of the research work (Chapters 4 and 5) includes four studies: In the first two (Chapter 4) an assessment method to identify driving profiles by means of 18 driving indices extracted from the HRT is proposed. In the first study reported in Chapter 4, two profiles emerged and were compared in terms of self-reported aberrant driving behaviors, whereas, in the second study, three profiles were identified and compared in terms of sensation seeking and decision making ability. In the last two studies (Chapter 5), the method previously proposed was employed to identify three “embryonic” driving profiles among adolescents without any license, also finding personality patterns that can predict the inclusion in the driving profiles; Overall, the results of this second branch indicated that the HRT can be employed as a tool to assess different driving profiles among low-experienced and totally inexperienced drivers and that these profiles show consistent links with the personality traits that are usually reported as predictors of risky driving behaviors. Finally, a first evaluation of the effectiveness of the training on the simulator in terms of on-road accidents and near misses (follow up) is reported, by means of an exploratory comparison of the self-reported real road performance of a trained and an untrained group of adolescents. The follow up study showed that, in a 12-months post-training period, participants reported less frequent bike crashes independently of the driving profile in which they were included. Moreover, an exploratory analysis showed that a subgroup of the trained participants, compared to an untrained control group, matched for age, gender and on-road exposure, reported less frequent bike crashes. Overall, the results of the present Ph.D. project suggest that: (1) a training of at least two sessions on the HRT simulator leads to a reduction in the risky and to an increase in the safe driving behaviors, in novice and inexperienced road users; (2) these changes are supposed to be due to an improvement in participants’ hazard perception, as attested by the psychophysiological changes in the participants who drove the simulator; (3) a one-session training on the HRT allows to have enough available resources to monitor the environment and to shift attention toward salient stimuli; (4) the behavioral improvements in the simulator seem to generalize to some degree to the real world, as attested by the reduction in the bike crashes for the trained participants and by the comparison of their crash rate with that of a matched control group; (5) the HRT can be employed as a tool to identify different driving profiles in people with different degrees of road exposure. The training and the assessment protocol presented in this research work can be considered useful tools to cope with the inexperienced road users crash rate issue, from a preventive perspective. The possibility, provided by the simulator, to test and train drivers in a safe and simulated environment, may lead to the development of ad hoc protocols, even customized on the basis of the participant’s driving profile and personality traits. As a result, this opportunity may provide a crucial contribution in preventing road crashes.