Database Reference
In-Depth Information
ing, for example, filtering or simple composition of primitive events, yet
can not be used for answering complex queries.
There is much work on complex RFID event processing, and represen-
tative ones include SASE [42] and RCEDA [73]. Next we will describe in
detail on the two frameworks in terms of event specification languages,
event detection models, event detection methods and their optimiza-
tions.
3.1 RFID Events
An RFID event is an occurrence of interest in time, and it can be
either a primitive event or a complex event.
A
Primitive RFID event
(also as atomic RFID event) is an event
generated during the interaction between a reader and a tagged object.
A primitive event is simple-semantic and represented as a triplet with the
form of
observation(r, o, t)
,where
r
represents the reader EPC,
o
represents the object EPC and
t
represents the observation timestamp.
The Electronic Product Code is an industry standard that defines unique
code to identify an object around the world. The unique identification
of each tagged RFID object through EPC code provides more semantics
for RFID events.
For example,
observation(r1, o1, t1)
represents an event gener-
ated for an object with EPC
o1
fromareaderwithEPC
r1
at time
t1
.
Primitive events are instantaneous, i.e., given any primitive event e, its
staring time equals to its ending time. Primitive events are also atomic:
a primitive event either happens completely or does not happen at all.
A
complex RFID event
or composite RFID event is usually defined
by applying event constructors to its constituent events, which are either
primitive events or other complex events. There are two types of RFID
event constructors: non-temporal and temporal, and the latter contains
order, temporal constraints, or both. While complex events defined with
non-temporal event constructors can be detected without considering the
orders among constituent events, complex events defined with temporal
event constructors cannot be detected without checking the orders, or
other temporal constraints among their constituent events, or both.
For example, shoplifting can be represented as a complex RFID event:
an item was picked at a shelf and then taken out of the store without
being checked out. This complex RFID event consists of three primitive
RFID events: two occurrences of the tagged item being detected at a
shelf and the exit respectively, and the non-occurrence of the item being
read by any check-out reader in-between.
