You say that "Several studies have suggested that at least part of this problem is due to insufficiently developed driving skills." Could you share what studies these would be?
Instructional support to drivers
By: M. J. Kuiken
Date: Monday, 18. February 2008
This is a summary of a Ph.D. thesis completed at Haren: Traffic Research Centre, University of Groningen
Summary
The main topic of this thesis was the use of feedback to efficiently reduce the occurrence of driving errors that might endanger road participants. The thesis initially focused on the position of inexperienced drivers. After a relatively short training period, newly qualified drivers are expected to drive safely without further supervision or guidance. Accident analyses demonstrate that young and inexperienced drivers are particularly vulnerable in traffic. The accident rate for this group is considerably higher than for more experienced drivers. Several studies have suggested that at least part of this problem is due to insufficiently developed driving skills. Feedback from the traffic environment is generally poor (McKenna 1982; Brown, 1989; Groeger 1990). Consequently, drivers easily fail to notice a discrepancy between their intended actions and realized actions. Inappropriate behaviour can thus become a habitual part of drivers' performance. Hence maintaining or improving safe driving procedures after drivers have qualified will require normative feedback. In this thesis we have investigated the effects of extrinsic feedback to drivers in situations in which they have been found to make errors, by means of an in-vehicle feedback system.
Error tendencies of inexperienced drivers were investigated using a postal questionnaire on behavioural antecedents of accidents. The rationale for using self-reports was that accident reports usually generate insufficient information about accident causes. This study gave some insight into drivers' critical behaviour in specific (accident) situations, which can serve to provide guidelines for the provision of extrinsic feedback on aspects of the driving task. Speed behaviour in curves was selected as the target behaviour of this study.
Although the survey revealed very little difference between the errors made by inexperienced drivers and those made by experienced drivers, accident statistics and additional studies suggest that excessive speed in curves seems to be related to drivers' level of experience and age. To investigate this issue further the driving performance of inexperienced drivers was compared to more experienced drivers. The methodology used, involved measuring drivers' speed when they were negotiating curves and straight road sections in an instrumented vehicle. However, once more, no clear differences between experience groups were found.
This led to the conclusion that behavioural errors of individual drivers need to be addressed independent of their level of experience, and that support by means of feedback needs to be adapted to the individual driver. Within the larger project of which this study formed part (i.e. the GIDS (project), a detailed normative task analysis was made and an on-line access to this database was made available for the use of our study. Basic notions were that drivers are goal-directed and that they are likely to regularly make mistakes or to fail to solve problems. Instead of providing feedback based on an analysis of error tendencies of different groups of drivers, feedback would need to be generated on the basis of real-time comparison between the actual performance of individual drivers and the normative performance. It was expected that such personalized feedback would facilitate the acquisition of safe, i.e. normative, driving skills by reducing specific errors.
Two experiments were carried out to assess the relationship between the personalized feedback and driving performance. The studies were performed in the driving simulator. Drivers received feedback on speed errors in curves with different radii, and/or at intersections. Errors were defined as deviations from the normative performance rules concerning speed behaviour in these situations. One of the main findings was that, compared to drivers who did not receive extrinsic performance feedback, drivers made considerably fewer and smaller speed errors when given feedback. These results suggest that performance support modifies driver behaviour. The findings also showed that feedback in supported situations transferred to unsupported but similar situations, i.e. other curves.
Another question in the experimental studies concerned the effects of various feedback characteristics, such as the timing and the modality of the feedback. The experiments investigated the effects of precise scheduling of feedback messages, and of verbal and haptic messages. What emerged most clearly from the findings is that both verbal and haptic performance feedback seem capable of conveying information to drivers on the basis of which the alter their behaviour. On the basis of the research reported here, it is clear that instructional support, which is properly suited to the traffic scenario in question, can exert a very substantial positive influence on driver performance. The general conclusion of this thesis is that in-vehicle feedback systems that monitor the output of the driver continually and compare it with the output that would be expected from the standpoint of safety, may play an important role in the future in the refinement and maintenance of certain behaviours. The present range of events and maneuvers over which the first generation prototype co-driver and tutoring systems are able to give support, is very limited. The next step is to develop the knowledge to include many more tasks and traffic situations that prevail in driving. Advancements in these areas will determine the rate of progress both of support systems in general and of feedback systems in particular.
Further comments to this article have been disabled.
All Comments (1)
Showing 1 - 1 comments
Michelle Rowell,
