Advances in sensor technology and wireless communication are enabling the real-time (or near real-tme) collection of information about the physical world. This raises many interesting possibilities as well as challenges for acquisition and visual presentation of this data. We start with a fully interactive, 3D model of a real-world venue. A user “flies” through the model and relevant sensor data, which impacts the rendering of the user’s current field of view must be fused with the static model in a timely fashion. One application scenario is emergency response in which sensors could provide information on the location of various resources, as well as environmental conditions as a function of location. The illustrations included below show one application in which service vehicles on the UCLA campus were equipped with GPS and wireless transceivers and the location tracking data was used to show in near real-time the actual location of the vehicles to a user while flying the 3D model of the campus.

Technical challenges include dealing with potentially thousands of sensors as well as multiple interactive users. Not only is the sensor data changing with time, but also the portion of the data that is relevant to any particular user is determined by their fields of view which are changing dynamically as they navigate the model. In an environment of limited bandwidth, getting just the right data to the right user and in a timely fashion is part of the challenge. In addition there are issues regarding the precision of the data required. Consider the application of following emergency vehicles in an area such as the university campus. When the user is “flying” at street level, relatively few vehicles may be in view but the precision and frequency of reporting their location needs to be high to obtain required accuracy. However, if the user flies up to view the overall campus from 1000 feet, all vehicles on campus will be in view and all their locations are needed, but with less precision and lower reporting frequency.