Safety view management for augmented reality based on MapReduce strategy on multi-core processors

Augmented reality (AR) can enrich intelligent transportation systems by overlaying virtual traffic information on user´s view. Pedestrians and drivers using AR applications can read traffic information and keep their eye on roads at the same time without distracting attentions. However, any information displayed on device screens may overlay on moving objects on roads and obstruct user´s view which causes traffic safety problems. This paper proposes a mechanism called safety view management to display information on AR applications in a safer way. When display labels or annotations on a device screen, the safety view management mechanism provides a safety region which does not overlap moving objects on the screen to increase driving safety. To efficiently detect safety region on a screen, the view management mechanism adopts MapReduce to analyze moving objects in parallel way on multi-core processors. Performance evaluations show that the proposed mechanism can efficiently calculate a safety region on screens and keeps drivers and passengers´ remain safety view in smart cars.