Methodology for the Evaluation of Safety Interventions

Abstract

In recent decades, automotive telematics and driver monitoring systems have been introduced in the industry in order to provide real-time and post-trip interventions and feedback to the driver. A few driver monitoring technologies and platforms have been used to record driving performance, focus on key risk indicators and provide safety interventions. Within that group of tools, interventions have been indicated to significantly enhance driving behavior and road safety. The purpose of the current study is to provide a methodology for safety intervention evaluation in order to keep driver behavior within acceptable boundaries of safe operation (i.e. Safety Tolerance Zone). To that aim, the most appropriate assessment variables from the i-DREAMS platform, related to the logic model of change were identified and some recommendations for the i-DREAMS project were provided. In order for the methodology to be designed, past experience on similar projects was exploited in order to derive a list of methods, indicators, utilized Key Performance Indicators (KPIs) and evaluation criteria mostly suitable for evaluating the project’s safety interventions. Three different methods (i.e. before-after analysis, case-control trials and questionnaires) were identified and therefore, the evaluation was conducted in terms of the outcomes proposed in the logic model of change. Results from literature review indicated that safety promoting goals and performance objectives had the greatest effect on the assessment of interventions. Driver behavior indicators, such as speeding, harsh acceleration or braking had the strongest impact on the interventions evaluation, while driver related characteristics, such as distraction, stress, fatigue, drowsiness and attention appeared to have lower impact. Taking into account the experimental studies, the design of a customized feedback strategy will assist in performing the appropriate evaluation of interventions needed for the improvement of driver behavior.