Evaluation of the Effectiveness of Volvo’s Pedestrian Detection System Based on Selected Real-Life Fatal Pedestrian Accidents

Abstract

The objective of this work is to test freely available system for active pedestrian protection. Tests are based on real fatal accidents that happened in the past with passenger cars that were not equipped with active safety systems. Tests have been conducted in order to evaluate what the real benefit of the active safety system is, and not to gain only a methodological prediction. The testing procedure was the first independent testing in the world which was based on real fatal pedestrian accidents. The aim of the tests is to evaluate the effectiveness of Volvo pedestrian detection system. An in-depth accident database contains about 300 fatal pedestrian traffic accidents in urban areas. Eighteen cases of pedestrians hit by the front end of the passenger vehicle were extracted from this database. Cases covering an average traffic scenario have been reconstructed to obtain detailed model situations for testing. Simulations of accidents have been made in PC Crash 10.0 using a multibody object and a mesh model of vehicles. Active safety testing scenario was built on the basis of reconstructed accidents with Volvo V40 cc and a new dummy simulating a pedestrian. Before the tests the dummy was evaluated in anechoic room to gain required radar reflection properties which would be the same as those of a human body. The movement of the dummy was driven by the autonomous ultraflat overrunable robot (UFO) for experimental advanced driver assistance systems (ADAS) testing and synchronized with Volvo motion by global positioning system (D-GPS) with high accuracy of motion.