Akademik Veri Yönetim Sistemi
11th ASME Biennial Conference on Engineering Systems Design and Analysis, (ESDA 2012), Nantes, Fransa, 2 - 04 Temmuz 2012, ss.803-811
One of the most challenging factors in the development of autonomous vehicles and advanced driver assistance systems is the imitation of an expert driver system which is the observer and interpreter of the technical system in the related driving scenario. To achieve an expert human-like situational understanding and decision making may be an important feature to fulfill the necessary active safety requirements. In this paper, an exploratory study on a multimodal adaptive driver assistance system is presented. The main goal is to determine the human driver's attention and authority level in a cognitive model and to trigger the timely warnings according to his/her driving intents and driving skills with respect to the possible driving situation and hazard scenarios. In the previous studies, a fairly restrictive vision-based driver assistance system has been deployed to detect lane departure, blind-spot and to monitor following distance, headway time. This vision-based driver assistance system considers the driver's driving performance metric sampled during the longitudinal and lateral vehicle control tasks as well as the processed information about the surrounding traffic environment consisting of the interactions with the other vehicles and the road situations. The presented active safety system models the driving task in a cognitive architecture and assesses the cognition of the human driver by modeling the situation awareness of the driver by using fuzzy sets. Each fuzzy set simply. represents the expert driver's perception in both of the longitudinal and lateral traffic. The presented system evaluates the driver's driving skills and attention level by comparing the expert and human driver's reactions suited in a finite set of decision and maneuvering task. In case of hazard analysis, the system triggers timely warnings pointing the driver's attention at the lateral or longitudinal maneuvering tasks depending on the interpreted situation. Introductory experiments are performed with a limited number of participants, the test driving data including the driver's perception and reaction to the surrounding vehicles and traffic situations are collected by the use of a vehicle simulator. And the presented multimodal adaptive driver assistance system is evaluated by the simulator. The preliminary results seem to be promising.