Table 3. Tests scenarios and related system reaction. 
Scenario Description  System reaction 
0  Standard procedure (see text)  Logging of all the operations  
1  An unknown user tries to access the 
document room 
Alarm generation 
2  A known user tries to access the room 
without an authorization 
Alarm generation. User ID is logged. 
3  Authorized user tries to enter the 
document room in the wrong time 
Alarm generation. User ID is logged. 
4  The authorized user opens the wrong 
shelf 
Alarm generation. User ID and shelf 
door coordinates are logged. 
5  The authorized user takes the wrong 
document 
Warning generation. User and 
document IDs are logged. 
6  The authorized user replaces a 
document in the wrong shelf. 
Warning generation and indication 
of the correct document position. 
User and document IDs, and 
document position are logged. 
7  The authorized user replace an 
incomplete document. 
Alarm generation and indication of 
the missing pages. User, document 
and missing pages IDs are logged. 
8  The authorized tries to leave the 
archive when the archive is not 
complete (missing pages or 
documents)  
Alarm generation and indication of 
the problem. User and involved 
document IDs are logged. 
4  Conclusions and Future Work 
In this paper we have shown the feasibility of integrating RFID technologies and 
security devices for the reliable management of sensitive document archives. There 
are still some open issues that need to be further investigated. 
The first issue is related to the setup optimization, in particular for what concern 
the tag answering time. As mentioned in Section3, currently we have used the default 
setup (worst case scenario), with the tags answering continuously whenever 
illuminated. The used tags can be programmed in such a way to reduce the answering 
rate. For example, it would be possible to reduce by 4 the answering time of the 
illuminated tags; this is equivalent to reduce, theoretically, the active tag population to 
¼ of the total amount making possible to have four times more tagged items placed 
inside a cell of the bookshelf without affecting detection performances. Regarding 
setup optimization, we expect that detection performances improve if tags of adjacent 
document pages are not overlapped. Next activity will therefore investigate the 
detection performance by varying the tag answering time and by changing position of 
tags in adjacent document pages.  
Another issue regards the management of electronic document copies. In 
particular, there is the need to trace also document copies by logging who made the 
copy and especially to give only to authorized personnel the possibility to physically 
copy the document for which (s)he has the authorization. To solve this question, we 
are going to add to our prototype a commercial document scanner that will operate 
97