Progress with situation assessment and risk prediction in advanced driver assistance systems

Abstract

In the field of automotive safety, advanced driver assistance systems (ADAS) are receiving growing attention. Effective ADAS requires awareness of the actual driving situation, a reliable assessment of the risks, and making rapid decisions on assisting actions. This paper reviews the current progress in these complementing subfields. The goal is to explore and critically analyze the most promising technological solutions and system application concepts. In order to systematize our study, first a reasoning model is introduced. Then, a detailed study of the different situation and risk evaluation methods currently applied in automotive safety systems is presented. Our first observation has been that the general thinking about ADAS reflects a ‘perception, analysis, decision and action’ pattern. In addition, we observed that situation and risk assessment is typically restricted both by the number of factors considered, and by a limited consideration of drivers’ attitudinal behavior. Though a huge amount of research knowledge has been published, there are still several gaps in the knowledge related to understanding and handling complex driving situations. We will use the information gathered in this survey to determine the most critical factors for ADAS, to extend risk assessment with consideration of human individual characteristics, and to develop a driveradaptive reasoning model.