A SURVEY OF RFID DATA PROCESSING - Managing and Mining Sensor Data - page 357

locations may change over time, as they are moved from one place to
the other. Besides state changes, events are also generated in the inter-
actions, including observations , when EPC tags interact with readers,
and transacted items , when an object participates in a transaction.
The dynamic entity-relation model (DRER) is an extension of the
ER model. In the ER model, all entities and relationships are static or
current. In the RFID system, entities are static, but the relationships
between them are dynamic. Thus, the only addition to the traditional
ER model is the addition of a new kind of relationship, known as the dy-
namic relationship . There are two kinds of dynamic relationships, one of
which generates events, and the other generates state history. An event-
based dynamic relationship is associated with a single attribute known
as timestamp , which represents the time at which the event occurred.
On the other hand, a state-based dynamic relationship is associated with
two attributes tstart and tend corresponding to when the state started
and ended.
Thus, in the DRER model, we have three different static entities cor-
responding to sensor reader , object ,and location . In addition, an entity
called transaction may be used in order to represent business transac-
tions, though we omit it in the discussion here for simplicity. Each of
the static entities is associated with its own set of static entity tables.
State-based dynamic relationships correspond to sensor location , object
location ,and containment . We note that each of these relationships are
dynamic, and naturally have a starting and ending time. Event-based
dynamic relationships occur at a particular instant, and may correspond
to an observation , which is generated by a sensor reading an EPC tag.
The different static and dynamic tables in the DRER model, together
with their attributes are summarized below:

Entity Type Table Attributes

Sensor Static

SENSOR(sensor epc, name, description)

Object Static

OBJECT(object epc, name, description)

Location Static LOCATION(location id, name, owner)

Observation Dyn. OBSERVATION(sensor epc, value, timestamp)

Containment Dyn.

CONTAINMENT(epc, parent epc, tstart, tend)

Obj. Location Dyn.

OBJECTLOCATION(epc, location id, tstart, tend)

Sens. Location Dyn.

SENSORLOCATION(sensor epc, location id,
position, tstart, tend)

These tables can be used in conjunction with a variety of SQL queries
in order to resolve interesting aspects about the RFID objects. Some
examples [68] of such queries are as follows:

RFID Object Tracking Queries: The OBJECTLOCATION table carries