Safeguarding of AI based perception for autonomous driving
Nikolas Brasch, Mahdi Saleh
The aim of the project is to develop and investigate methods and measures for safeguarding AI-based functions and individual modules for automated driving. The aim is to make it easier to determine and assess the residual risk that cannot be ruled out when using a technology that appears opaque to humans and is therefore often referred to as a "black box". As part of a pipeline for vulnerable road user (VRU) protection the chair provides 3d object detection and human pose estimation methods, which will be analyzed and optimized with respect to their safety properties.
AID, ASTech, Audi, BMW, Bosch, BUW, Continental, DFKI, DLR, FZI, HCI, Hella, IAIS, Intel, Mackevision, Merantix, Objective, Opel, P3, QualityMinds, Valeo, Visteon, Volkswagen, ZF