Research On Warning Method Adaptive To Driver Drowsiness And Traffic Environment

Driver drowsiness resulting in deteriorated vehicle control is one of the majorcauses of road accidents. For this reason, developing systems that could issueappropriate warnings when he/she is not paying enough attention to the road is apromising way to avoid potentially hazardous situations and accidents caused bydrowsy driving. However, the degradation of drowsiness level may eventually diminishthe driver’sability of perception, estimate and response when an emergency isapproaching.lt is still an unsolved problem how to design a hazard warning systemadaptive to driver’s drowsiness level, and avoid the interference among differentwarnings at the same time. On the other hand, it is notoriously diiffcult to design ageneral warning system adaptive to all kind of drowsiness. Therefore，this paperconducted an adaptive warning strategy and presents a warning method which isadaptive to the combination of driver’s different drowsiness level and traiffc hazard.For designing the adaptive warning strategy, the driver’s perception of traiffchazard in different drowsiness level was studied in detail. An adjustment method ofwarning timing based on driver’s perceived risk was proposed, which achieved thetraiffc hazard warning timing adaptive to driver’s difference of drowsiness level. Thenthe priority of driver’s drowsy warning and traiffc hazard warning was evaluatedwhereby we provided an information management strategywhich avoided theinterference between different warning information.For the sake of warning effectiveness in different level of drowsiness, anevaluation system was conducted based on information processing mechanism ofhuman. Then the paper evaluated the warning effects of each speciifc single warning,such as auditory, visual and tactile signals for the driver in different drowsiness leveland traiffc hazard situation. Furthermore, the inherent law of overlying effect ofmultiple warning was excavated from the point view of perception mechanism, and thebasic pattern of multiple warning effects was achieved, which lead to set up a designmethod for adaptive multiple warning. On this basis, the multiple warning adaptive todriver’s different drowsiness and traiffc hazard was ifnally formed. The validation experiments for the adaptive multi-warning method were conductedin drive simulator and real road as well. The results showed that the warning methodproposed in this paper could provide a well-timed and pass muster warning for thedriver in different drowsiness level and the overall effectiveness was signiifcantlyimproved than any single warning.