In this task, alternative methods for tracking the user's point of view in an AR application will be explored with results including a set of recommendations for the final implementation.

8.1.05

Searched for different tracking options and costs/benefits. Most systems are expensive (~$10,000+). Remembering that Conspiritus had used a tracker of some sort, we go to their website and try to send them an email to find out what they went with and why.

8.2.05

Settled on a few different options. Need further investigation. Low cost option with high range but possibly bad accuracy and tethered to field antenna is DC magnetic tracker, SpacePad used for many VR games. ~$1000

8.10.05

We investigated another possibility discussed in our original project proposal--markerless, feature-based tracking. We downloaded and test ran two demo programs from the CV Lab in Lausanne. One tracks facial features and adds AR glasses and a mustache. This tracked X, Y, & Z position information but seemed buggy and would not track pitch, roll, and yaw orientation information unlike the code they're using in their own labs. The other demo would only track the specific cover of a book they had programmed. It appeared it might be possible to change this tracking data, but it used a very unusual format that didn't appear to be very often used or well-documented elsewhere on the web.

8.23.05 - 8.24.05

Continue web search and bookmarking possible products which may serve as a sensor-based tracker instead of, or in addition to, optical fiducial markers.

8.25.05

To make sense of the overwhelming number of options Stephen begins work on a spreadsheet to compare the different features, options, and drawbacks of each viable product.

8.26.05

Finally get ahold of Steve Kearsley at the Exploratorium. He provided a few tracking ideas to build upon with regard to use of the Teleo boards and how we might be able to use them because they have Actionscript drivers which may be accessible from Macromedia's Director which DART uses.

9.6.05

Searched the web for more "low-level" sensor technologies that we could use in conjunction with Teleo boards if we decide to make our own motion tracking sensor and program the transformations ourselves in DART. Found a good option with respect to performance and price--6DOF IMU (Inertial Measuring Unit), ~$450.

9.10.05

Song discovered some information indicating that 3D panels defining markers in the shape of a cube/room is possible. Though we had an idea on this earlier this summer, we didn't know where to begin. Song's discovery led to Stephen finding resources explaining how to perform the necessary mathematical matrix calculations required to define a matrix panel of virtually any shape (2D or 3D) for strategic placement of world coordinate markers within the user's environment.