Key points are not available for this paper at this time.
The incorporation of Advance Driver Assistance Systems (ADAS) is a growing trend in active safety systems. The current outlook of passive ADAS that, exclusively warning the driver, presents a reactive interaction model in which each system acts independently. The increasing of the amount of in-vehicle passive ADASs arises an interoperability issue against conflict situations that could affect driver's decision-making process negatively, diminishing their attention level. This paper proposes an architecture based on the multi-agent paradigm for designing driver-centered ADASs that operates through the data fusion. The principal goal is to design a hierarchical structure which can manage the knowledge acquisition process of all aspects involved in the driving scene such as the environment as well as the driver's behavior and state, providing support for building and testing reasoning models. An experimental method was performed to verify the feasibility of the proposed approach using the deployment of a warning ADAS which involves the interaction of several developed systems. Besides, the staging of one set of specific hazardous driving situations was designed to conduct an experimental assay with ten drivers in a driving simulation system. Regarding evaluation measurements, the analysis of the reaction time performed by the drivers and user questionnaires collected after each experimental session has shown promising results.
Sipele et al. (Sat,) studied this question.