Filtering Redundant Data from RFID Data Streams

Table 1

Summary of existing duplicate RFID data filtering approaches.

Approaches	Weaknesses
DLTBF [10]	Needs to compute the least loaded bucket before insertion. Also, it uses a number of hash functions.
CBF [9]	Filters duplicate readings only at a reader level. It also uses a number of hash functions.
BF [12]	The filter is easily becoming full since it does not allow deletion and therefore generates a higher rate of false positive.
TBF, TIBF [7]	Not memory efficient because multiple time counters are needed to store the reading time.
Sliding windows [19]	Inefficient since it has to scan along the sliding window every time new reading is coming in which it scans almost the same data. Besides that, sliding windows can overlap but disjoint the first items.
Landmark windows [19]	Size of the window can be very big to provide more accommodation and accurate results. Besides that landmark windows can overlap in any way.
CCBF [8]	Delay process because when it gets to identify probability of duplicate, it has to search the double-linked list to check whether the intersection of all time intervals corresponding to each hash function is empty so that the tag did not arrive within time. Just related dynamic setting with the object’s arrival rate and not the departure rate which indicates readings that can be removed from the filter.