Technology and Low Fidelity Prototype

Identification for Scheduling and Sorting

RFID is neither precise nor inexpensive enough for our prototype. The current sensor is too large to use as a single cell grid, and cannot determine the position of a cell on the platform to use it for the entire grid. Purchasing more sensors would be too costly.

Computer vision requires a camera, computer and software. The lab, the internet, and our programmers can provide those. Furthermore, computer vision can identify the presence of individual objects in view if they have been previously registered. It can also determine to location and orientation of the objects on a single plane (perpendicular to the viewing angle).

Two methods may be used to identify the pills: color or registered patterns. Color would allow for low fidelity prototypes to look simpler, because the "pills" could look like actual pills. Registered patterns however, may be easier as both TopCodes and AR Toolkit employ this method and both are robust enough for our purposes.

We are currently looking into adjusting AR Toolkit to identify colors, but are prepared to use patterns for the prototype.

Initial Prototypes

Made from Crayola clay and some other white modeling material. Includes schedule grid, storage containers, pills, and possible travel pill-box designs