Since driving in haste leads to a considerable number of traffic accidents around the world and since research in the area is rather limited both in terms of quantity (studies are rather limited in number) and quality (most of the existing studies are questionnaire-based and the others have limited sample sizes), the overall objective of this PhD research was to identify indicators with sufficient discriminative power for the recognition of driving in haste that could eventually be used in the development of algorithms for the detection of this state in real driving conditions More specifically, the aim was to further understand the phenomenon of haste in order to: (i) induce this state for the execution of a study using a driving simulator, as well as to (ii) identify, based on studies reported by other authors and knowledge derived from experts in the fields of driving, behavioural psychology and human medicine via focus group studies, possible observable manifestations or indicators of driving in haste, and finally (iii) empirically assess the discriminative power of the indicators identified for the recognition of driving in haste by means of the design and execution of an experiment using a driving simulator. From the above, the main assumption for the current study was that the behaviour of the driver in a haste situation would be different from his/her usual behaviour in a normal driving situation and that the emerging changes in the behaviour would propagate from the driver through the car towards the environment. That is, the behavioural state of the driver would be reflected not only on the human body, but it would also have observable influences on the behaviour of the car, the interaction of the driver with the car, and the interaction of the car with the environment. Moreover, the global hypothesis for the current study was that there existed at least one indicator that was completely independent from the driving context and completely insensitive to differences between drivers. In other words, prior to the conduct of the study, it was believed that it would be possible to find one indicator that would have sufficient discriminative power for the recognition of driving in haste independent of whether the driver was one or another, and whether he/she was driving on a freeway or approaching an intersection with traffic crossing ahead. A secondary hypothesis for the current study was derived from the fact that, in the overall process of driving a car, the behaviour of the car (i.e. vehicle information such as speed, acceleration, etc.) is a reaction to the drivers’ behaviour and, thus, indirect driving parameters are responses to direct driver-related actions. Consequently, since both indirect and direct measures provide similar information, it was expected that indicators closer to the source, that is indicators coming from the physiology of the driver and his/her interaction with the vehicle controls, were best for identifying driving in haste. In fact, it was initially expected that indicators closer to the source alone were already enough for the intended purpose. Another hypothesis was related to the induction of haste in drivers in a laboratory setting using a driving simulator. The premise was that driving in haste, with the stressor being time-related, could be induced using the time pressure construct defined as a time constraint plus a motivation to perform the driving task in the given time. In other words, it was expected that haste resulting from a time-related stressor could be intentionally prompted by reducing the time available to complete the task as long as the participant was motivated to do so. (Author/publisher)
Abstract