Decision support issues in automated driving systems

Abstract

Machine learning and computational processing have advanced such that automated driving systems (ADSs) are no longer a distant reality. Many automobile manufacturers have developed prototypes; however, there exist numerous decision support issues requiring resolution to ensure mass ADS adoption. In the coming decades, it is likely that production ADSs will only be partially autonomous. Such ADSs operate within predetermined conditions and require driver intervention when they are violated. Since forecasts of their 20-year market penetration are relatively low, ADSs will likely operate in heterogeneous traffic characterized by vehicles of varying autonomy levels. Under these conditions, effective decision support must consider intangible, subjective, and emotional factors as well as influences of human cognition; otherwise, the ADS risks driver distrust and unsatisfactory performance based on an incomplete understanding of its environment. We survey the literature relevant to these issues, identify open problems, and propose research directions for their resolution.

Managing Driving Modes in Automated Driving Systems

“…Provided that recent estimates indicate semiautomated level-3 and -4 ADSs will constitute 25% of the global market by 2040, traffic in the coming decades will be a heterogeneous mix of manual and automated vehicles. Therefore, until level-5 vehicles predominantly populate global roadways, RtI decisions and their management will remain a crucial safety-related issue (Caballero, Rios Insua, and Banks 2021). However, relatively few studies focus on the RtI decision.…”

Some statistical challenges in automated driving systems

“…Myriad exogenous factors also affect society's willingness to adopt ADS technology. Figure 1, adapted from Caballero et al, 9 integrates anticipated impacts of ADS technology on society with factors affecting its adoption. This network illustrates the interconnected nature of numerous societal considerations with ADS adoption.…”

Safe Vehicle Trajectory Planning in an Autonomous Decision Support Framework for Emergency Situations

“…The autonomous mode uses the planning algorithm developed for the Co-Pilot, while the emergency mode uses a new and specific algorithm for emergency situations, detailed in Section 3. The early warning module was added to try to prevent emergency situations whenever possible, especially in "human driving" mode (when the Co-Pilot is not used), and is detailed in [68]. In the Trustonomy project, the Decision-support module also interacts with the HMI (Human-Machine Interface) and DIPA (Driver Intervention Performance Assessment) modules, but these interactions are not detailed in this paper.…”

“…When a risk is higher than the warning threshold, the associated warning is sent to the driver via the HMI, whereas the emergency threshold is used directly by the Mode manager. This component is presented in details in [68].…”