Research work

RFID could be used to determine presence or absence of a physical component, but I would suggest you look into a visual system or computer vision platform of some kind to determine position. Using RFID is not ideal for precisely locating something... in fact, using RF in general is not ideal.

An example of why not is the E911 phone location service. Before GPS was common in phones, cell towers were fitted with special gear and special antennas used to triangulate phone position as accurately as possible. They use not just RSSI signal strength, but AoA (angle of arrival) and TDoA (time & direction of arrival) in order to calculate the most accurate position possible, yet a phone's location can only be narrowed down to "a location somewhere within a 100 meter radius of X". That means a phone can only be said to be somewhere inside a giant circle 200 meters across. Using RSSI only, which is the only signal data coming from RFID systems out there, positioning accuracy deteriorates significantly.

For OpenBeacon, there are several videos of the OpenBeacon location tracking project that show avatars moving smoothly across the screen, which seems to indicate a very accurate positioning system was built using OpenBeacon. The fact is, only very crude positioning data was able to be obtained using OpenBeacon and some complex math functions were used to determine direction of travel and smooth things way out. When people stopped moving, the function lost it's power and actual position was pretty much a rough guess.

Let's talk now about storing data on the tag. Is there a reason for keeping your data on tags which may or may not be present? Why not keep each unique tag ID tied to a record in a local database and just cross-reference tag IDs with records in the database? That way you could also know which tags were not present, and know what physical objects those absent tags were affixed to. Often times there is little or no reason to actually store data on tags. The only valid reason I can think of is if you needed to transport data to sites or systems that were not able to directly access a common central database. Any time a tagged object has to communicate data with sites or systems that are not related (product moving through manufacture to retail to consumer for example), those types of situations may warrant writing data to a tag.

Now let's get back to vision systems. There is a lot of work being done in the VR/visual space with the Microsoft Kinect. You could probably do away with RFID all together and mark the physical objects with 2D barcodes like QR codes or a datamatrix code and be able to identify and position the object properly. By knowing what the object is, it's dimensions, and knowing the barcode is positioned X by Y inches from the top left corner of the object (or in the center), you could determine edges and position. You could also use registration marks to determine angle of view. Actually, I think the Kinect already throws out an IR grid pattern that the IR camera can see, so angle might already be figured out for your by the Kinect.