The level of sophistication exhibited by RFID tags is not only affecting their financial cost but also their ability to provide (extensive) cryptographic functionalities. It follows that low-cost tags offer no real control access. To tackle this issue, we propose an anomaly detection system which attempts to identify deviations from the normal behavior. In practice, subjects are equipped with RFID tags in order to be constantly located. Then, a user's profile is built, relying on the Kohonen's maps that constitute an efficient way for automatically categorizing and further compare the tag behavior against the normal user’s behavior (as expressed in the user’s profile).