Road safety is a complex issue influenced by a wide range of factors, including road conditions, vehicle characteristics and driver behaviour. The aim of this PhD thesis was to improve driver Safety Tolerance Zone (STZ) through a holistic analysis of road, vehicle and behavioural risk factors. More specifically, the impact of task complexity and coping capacity on crash risk was examined. Towards that end, data from 190 drivers who participated in a large on-road and simulator driving experiment were exploited. An innovative methodology, consisting of both statistical and machine learning analyses, has been developed and implemented, including Generalized Linear Models (GLMs), Structural Equation Models (SEMs), Neural Networks (NNs), Decision Trees (DTs), Random Forests (RFs) and k-Nearest Neighbors (kNNs). SEMs demonstrated that task complexity was positively correlated with risk, indicating that driving during night-time or in adverse weather conditions can exacerbate the challenges posed by complex tasks, further increasing the likelihood of crashes. Conversely, coping capacity was negatively correlated with risk, indicating that drivers with higher coping capacity are better equipped to handle challenging driving situations. Results indicated that RF models outperformed the DT and kNN models across all metrics, making it the most effective for predicting speeding and headway, with overall accuracy up to 90%. NNs demonstrated that the level of STZ can be predicted with an exceptional accuracy of up to 89.8%. Lastly, it was demonstrated that simulator experiments proved to be the most suitable for predicting STZ levels and naturalistic data, without simulator validation, may lack the controlled conditions necessary for thoroughly evaluating complex interactions. Overall, it was observed that both real-time and post-trip interventions had a positive effect on driving behaviour, as drivers managed to improve their performance. The integrated treatment of task complexity, coping capacity and risk can improve behaviour and safety of all travellers, through the unobtrusive and seamless monitoring of behaviour.
| ID | at9 |
| Full Text | |
| Tags |
