Search Results

Driving is a very complex activity which requires all the attention of the driver. Nowadays, the human factor is related in 90% of the accidents and the distraction is one of the most considerable factors of influence, specially the somnolence. The objective of this project is to create a confident somnolence detector. In order to obtain essential data from the drivers while they fall asleep to calibrate the algorithm, we needed to create a system which allows the driver to handle the car, but detects when the conduction is not correct and takes control to keep safe the driver and the car. It was extremely important that the system must be autonomous, and the driver must be alone in the car, so as not to distort the experiment. As a result, we will obtain a semi-autonomous non-intrusive system which makes impossible to have an accident in our platform where a track is circumscribed.

Around 15% of traffic accident casualties in Europe are pedestrians. To date, most of the studies carried out only provide statistical information on the problem and little study in-depth the countermeasures which might correct it. There are many studies concerning pedestrian protection, which can be grouped into ‘pedestrian modelling’, ‘biomechanical limits for pedestrians’ and ‘statistical analysis for pedestrian accidents’. Despite these studies, there is no predictive analysis of the benefits of pedestrian protection systems based on their intrinsic capabilities applied to a real accident sample. This paper describes a methodology for the evaluation of pedestrian protection systems based on the analysis of a wide sample of urban pedestrian accidents. All of them are analysed in-depth and reconstructed with PC-Crash. The effects of the frontal structure of the vehicles and several active systems, such as BAS and Pedestrian Detection Systems, are evaluated.