reliability improvement for an rfid based psychiatric patient localization
TRANSCRIPT
Reliability
improvement for an
RFID-based psychiatric
patient localization
system
Reference
Chieh-Ling Huang, Pau-Choo Chung, Ming-Hua Tsai,
Yen-Kuang Yang, Yu-Chia Hsu Reliability improvement
for an RFID-based psychiatric patient localization
system IEEE Computer Communications 31 (2008)
2039–2048
2
Outline
Introduction
System overview
Reliability improvement with field generator scheduling
Experiments
Conclusion
3
Introduction
Psychiatric patients often cannot control their actions, occasionally resulting in dangerous behavior
RFID technology has been utilized in various applications, including supply chain management, entry and exit control
Localizing moving objects, e.g., freight localization or human localization is a challenging and relatively unexplored task
presents a novel graph coloring with merging and deletion (GCMD) algorithm
4
System overview
The Department of Psychiatry, National Cheng Kung University Hospital (NCKUH) uses an RFID-based psychiatric patient localization systemThe second floor serves as a clinic for psychiatric patients, and the third floor
is an activity area
Nurses : scheduling daily activities and providing basic care
Doctors : medical treatment
This RFID-based psychiatric patient localization system uses a ultra high frequency (UHF) long range tracker. The Field Generators operate on 433 MHz when triggering the Tag to respond, and the Tag replies to a Reader with 916 MHz signal once it is triggered
Psychiatric patients in the care center wear watch-like Tags, Tag transmits information, including Tag ID and Field Generator ID
5
System overview
6
Reliability improvement with
field generator scheduling
This system relies on the Tag correctly receiving signal from the Field Generators to estimate the Tag location
Two Field Generators with overlapping transmission ranges simultaneously issuing trigger signals to a Tag causes signal interferences in the overlapped region
This interference results in loss of signal and, therefore, decreases localization accuracy
Transform the relationships among all Field Generators into a graph
Vertex deletion and merging
Vertex coloring
Operation slot allocation
7
Reliability improvement with
field generator scheduling
8
Transform the relationships among
all Field Generators into a graph
The relationships among Field Generators are
transformed into an undirected graph G, whereas V and
E are sets of vertices and edges, respectively. Where V
and E are derived based on the Field Generators and
their signal region overlapping situations, respectively
9
Vertex deletion
When the range of one Field Generator, FGx, is
completely covered by other Field Generators, the
function of FGx can be replaced by a combination of
these other Field Generators
10
Vertex merging
The entire overlapping region is covered by a union of
Field Generators, so other Field Generators can cover
the overlapping region
Consequently, the two vertices associated with the two
Field Generators can be merged, and no special care is
required to avoid signal interference from the two Field
Generators
11
Vertex coloring
A coloring algorithm is applied to the trimmed graph, in
which connected vertices are assigned different colors
The Field Generators with the same color are assigned
to the same group, and, therefore, can transmit signals
simultaneously
Conversely, the Field Generators with different colors
are assigned to different groups; scheduling must be
applied to avoid signal conflict
12
Operation slot allocation
A weighted TDMA is applied to assign time slots for operation to each group
Consider that each partitioned group can occupy different levels of importance
Another consideration is the size of an area covered by a group of Field Generators
The importance factor for each group wi can be
approximated
The time slot ratio for each group
13
Experiments
Elucidating system performance
Fixed-points test
Route test
14
15
16
Elucidating system performance
17
Elucidating system performance
Tags
send out responses periodically (reciprocated regularly)
only when triggered by Field Generators
A patient’s location is computed based on the
communication range of the patient’s Tag within the
Field Generators with respect to ranges of reference
Tags
18
Elucidating system performance
More than two Tags are sending reports back to the
Reader simultaneously Repetitive transmission
Repetitive transmission times are set at 6 and the
associated lasting time is Trep
19
Elucidating system performance
Time in field (TIF) time: a Tag can be programmed with
a TIF Time (TTIF) that specifies the time duration before
the Tag can be triggered again
A Reader receives two consecutive reports from the same Tag.
How can the Reader determine whether the two reports are
issued due to two separate triggers, or whether the two reports
are due to a repetitive response trigger?
Another aim of TIF time is to prevent Tags from wasting
energy replying to the same trigger from a Field Generator
20
Elucidating system performance
Trep + TTIF is defined as one round; if one of the six responses in one round is received, this round is regarded as successfully received
lost rate of responses L as the total number of lost rounds divided by the total number of rounds that should trigger the Tag: r denotes the number of rounds that the Reader successfully received Tag’s reply signal and Trepresents time cost
In this system, it takes roughly three rounds for a patient to move from the building exit to the main gate. Under this scenario, we define response rate Rn as: n is the number of rounds – 3 in this case
21
Fixed-points test
22
positions 1–6 reside in single Field Generator range
positions 7–10 are located in the overlapping region of twoField Gen-erators
positions 11–14 are in the overlapped region for three Field Generators
position 15 is in the overlapping region of four Field Generators
people wearing Tags stand at each fixed position for 1 min
group1 is assigned 2 s for operation and group2 is assigned 1 s
Fixed-points test
23
Fixed-points test
24
Fixed-points test
25
Route test
5 routes
(a) Route1 is the path passing the 15 representative points
(b) Route2 is the path connecting with poor reliability in the
fixed-point test
(c) Route3 is the path connecting points with high reliability
(d) Route4 is the path of shortest distance from the building
exit to the main gate and
(e) Route5 is a route tracing through a region that is rarely
covered by routes (a–d)
26
27
(a) The route connecting 15 representative points
(b) The route connecting the lowest reliability points
(c) The route connecting the highest reliability points
(d) The route having the shortest path from exit to main gate
(e) The route tracing a region not tested in (a–d)
Route test
28
Route test
29
Route test
30
Route test
31
Experiments
position 3: The Field Generator has difficulty reaching
this sharp corner and the Tag cannot reach the Reader.
Thus, a Reader is added at position 10
Experimental results: response rate for position 3 improves
from 0% to 57.81% in the unscheduled original system,
and from 14.26% to 83.36% using GCMD scheduling
Transmission time slots should be based on group
importance
For group covering important regions or large areas
should be allocated increased time periods
32
Reliability comparison for original system and the
system with proposed GCMD algorithm
33
Conclusion
The RFID devices that are small and relatively cheap are
very appropriate for use in localizing psychiatric Patients
In this study, a GCMD scheduling model is utilized for
scheduling Field Generator transmissions in an RFID-based
psychiatric patient localization system, thereby reducing
interference caused by Field Generators located near one
another
Experimental results demonstrated that the system is highly
effective when using the proposed scheduling algorithm
34