Augmented Reality (AR) applications place virtual contents, such as markers, avatars, images or videos, in users’ views of their surroundings. Users of outdoor AR applications can use these while walking, riding a bike, or driving in a vehicle to, for example, navigate dense environments safely. Today’s AR frameworks lack comprehensive support for developing outdoor AR applications. This project will produce a suite of techniques that will enable advanced AR applications, especially those that permit multiple users to interact with virtual objects in the environment, and deliver continuously updated views in near real-time of moving objects in the environment. Every AR application involves three different entities: the environment, the user, and the virtual objects introduced by the AR application. The collaborative project, which brings together researchers from the University of Southern California and the University of Michigan, will systematically develop techniques to model the environment, determine human pose, and model virtual object interaction with the physical environment. It will explicitly focus on fast algorithms, resource-efficient realizations on mobile devices, and employ a judicious combination of latency hiding, offload, pre-computation and pre-fetch techniques to simplify the development of usable outdoor AR applications with rich functionality. Outdoor AR applications can potentially improve public safety, enhance education, promote public health