Analysis of Sensing Systems for the Detection of Pedestrian Impacts

In order to reduce pedestrian casualties in vehicle traffic accidents, safety countermeasures, for example the active hood system and external airbag system, have been proposed to protect pedestrians in vehicle collisions. Such a protection system must consist of a sensing system to detect the pedestrian impact and/or the potential impact. It is a key issue to verify the effectiveness of the detection system. The aim of this study is to evaluate the performance of a remote sensing system and a contact sensor for the detection of pedestrian impacts. In paper I, a remote sensing system was evaluated for the pedestrian detection in vehicle traffic environment. In this study, the car-pedestrian accident data were selected from the Swedish Traffic Accident Data Acquisition (STRADA), a database developed by collecting accident data from the police and hospitals. In this database, 9 scenarios of car-pedestrian accidents were defined according to pedestrian and car trajectories and accident locations. A total of 2,199 car-pedestrian impacts with identified accident scenarios were selected from the database and further analyzed in this study. Based upon the analysis, the two most common scenarios–accounting for 46.8 percent of all the cases–were identified as cars entering and leaving intersections colliding with pedestrians crossing the road. Based on this understanding, the knowledge of these two scenarios was developed in terms of the trajectories and speeds of pedestrians and cars: the qualitative pedestrian and car trajectories were described by the definition of the accident scenarios, while the quantitative speeds of pedestrians and cars were estimated based on pedestrian ages and road speed limits. By using both the qualitative and quantitative data, a mathematical model was presented with the hypothesis that all the pedestrians in both scenarios were not obstructed. The effectiveness of the sensing system for the detection of potential pedestrian collisions was then evaluated by using this model. In paper II, a new pedestrian air-tube contact sensor in the car bumper was developed at Autoliv Research due to that a traditional contact sensor can be unstable at varying temperatures. This paper aimed to evaluate the new contact sensor for the temperature-independent measurement of pedestrian impacts. First, the baseline FE bumper model of a production car was developed and validated according to the EuroNCAP lower legform impact tests performed on the car bumper. The results of the validation analysis were comparable with the test results, but they have exceeded the limits of the EEVC WG17 lower legform impact tests. The baseline bumper model was therefore considered as valid, but it required an enhanced safety performance. Based on the baseline model, an improved bumper model was subsequently developed to meet the EEVC requirements. Following the EEVC test methods, the legform impact simulations were conducted in the middle and on the left and right sides of the improved bumper model, resulting in the maximum legform responses of tibia accelerations from 117 to 139 g and knee bending angle from 7.7 to 9.3 degree, thereby meeting the EEVC requirements. Second, a FE human lower extremity model was developed; the baseline and improved bumper models were further evaluated using this human model. The improved model was shown to protect the human knee joint effectively, but the risks of tibia and fibula fractures were increased. Finally, the contact sensor was built into the improved bumper model and an analysis was conducted to evaluate the sensor performance. From paper I, it was concluded that the sensing system can detect almost all the pedestrians in the two scenarios, if the detective angle of the system is greater than 60 degrees. From paper II, it was determined that using a 25-mm instead of 50-mm air-tube sensor resulted in more stable sensor output at varying temperatures in the EEVC WG17 legform impacts; moreover the sensor was more sensitive to the different masses of the impact objects. The 25-mm air-tube sensor was thus a better choice for the contact sensor design