A combined comfort and safety-based approach to assess the performance of advanced driver assistance functions

Abstract Advanced Driver Assistance Systems (ADAS) play a fundamental role in improving the driving experience by enhancing both comfort and safety. Currently, there is a lack of objective methods to quantify ADAS performance comprehensively, especially in different emergency conditions such as avoidable and inevitable collision states. This study introduces a framework for evaluating ADAS performance in any condition of potential interaction between two or more vehicles. For impacts that are avoidable by the system’s intervention, assessments are based on comfort using vehicle acceleration (ISO 2631) and safety considering the minimum distance achieved between vehicles (clearance). In cases where the impact is inevitable, the performance indicator focuses on the injury risk for vehicle occupants associated with the collision and resulting from the ADAS intervention. To illustrate the application of this methodology, two case studies are presented, involving vehicles equipped and not equipped with ADAS. These cases respectively represent a real impact extracted from an in-depth accident database and a consumer program test for an Adaptive Cruise Control (ACC) function. By analyzing the behavior of autonomous steering and braking functions in a simulation environment as ADAS parameters are varied (i.e., the time between consecutive scans of the scenario by sensors and the time for full actuation of drive-by-wire systems), it becomes possible to compare performances of different types of intervention logic. This approach not only facilitates the identification of the performance of potentially available ADAS functions (as ACC) but also helps in highlighting the best logic applicable in a specific conflict. The retrieved information has the potential to indicate margins for improvement of ADAS already on the market, and simultaneously guide the development of next-generation ADAS with an increasing focus on autonomous driving and enhanced user technology acceptance.