Performance Characterization on Mobile Phones for Collaborative Augmented Reality (CAR) Applications

Collaborative Augmented Reality (CAR) systems allow multiple users to share a real world environment including computer generated images in real time. Currently, the hardware features of most mobile phones not only provide excellent multimedia services, but it also includes wireless network capabilities that provides a natural platform for CAR systems. However, the wide variety of these hardware features can have important effects on the performance of the mobile CAR applications. This paper presents the experimental characterization of CAR applications for mobile phones in regard to well-known performance metrics in distributed systems. Characterization results show that the most time consuming stage in a CAR application is the marker detection stage. Moreover, the rendering stage is decoupled on some devices. This decoupling process allows avoiding low refresh rate, facilitating the collaborative work. These results can be used as the basis for an efficient design of CAR systems and applications.