FFI - Fordonsstrategisk Forskning och Innovation

Novel methods and tools have been developed to determine active safety system performance using computers within the project “Verification Methods for Active Safety Functions”. Special attention has been given to include the non-ideal behavior of sensors, to deal with the large variation within the traffic environment where the system operates, and to develop methods to determine system robustness and performance. The project has delivered: - Method and tool to determine the sensor errors of camera based active safety system; - Method to determine the worst case performance of collision avoidance and mitigation systems and the corresponding traffic scenario as well as system robustness; - Method for formal verification of collision avoidance systems, and system robustness to measurement errors; - Method to predict the benefit of active safety functions applied to Volvo Cars’ “Full Auto-brake with Pedestrian Detection” based on real world fatal pedestrian accidents; - Method for sensitivity analysis and tuning of active safety system; - Method for reliable vehicle pose estimation using vision and single-track model; - Method to develop pedestrian detection algorithms using augmented training data; - One PhD with scientific publications; - Presentations at SAFER, Chalmers and international conferences