Primary safety systems have been developed for vehicles in order to detect a pedestrian and to avoid or mitigate an impact autonomously. This work aims to estimate the safety potential of six Active Pedestrian Safety Systems (APSS) from a sample of 100 real vehicle/pedestrian crashes provided by in‐depth crash investigation. The accident cases were first reconstructed by emulating the kinematics of the vehicle and the pedestrian. These simulations provided a comprehensive set of data describing the interaction between the vehicle and the pedestrian over a crash sequence. Then, four particular pre‐crash events on the timeline were selected as fields of interest with respect to performance characteristics of APSS. They correspond respectively to 2.5s before the impact, the instant when the pedestrian is visible (pedestrian steps into the field of view of the sensor), the last moment for the vehicle to brake in order to stop before impact and one second before the impact. For each of these instants and for each of the six selected APSS, it was evaluated if the systems could detect the pedestrian according to the different attributes of these systems. Results allow describing the required performance of an APSS and understanding the issues and challenges in pedestrian safety.
Keywords:
Pedestrian active safety, Primary safety systems, Pedestrian crashes, Crash reconstruction, Crash modeling