The present invention is related to the field of Wi-fi device localization in an indoor
environment, and more particularly related to an application of user dynamics through Wi-Fi
device localization in an indoor environment, where a user having the Wi-fi device is moving
within the indoor environment.
10 BACKGROUND OF THE INVENTION
Background description includes information that may be useful in understanding the present
invention. It is not an admission that any of the information provided herein is prior art or
relevant to the presently claimed invention, or that any publication specifically or implicitly
referenced is prior art.
15
Crowd movement inside a market arena has become an area of interest as it deals with business
management. Each individual in a crowd is different in their choice of commodities that they
want to purchase from a certain market arena, and therefore how they approach the market
arena with their needs vary considerably. Some individuals in the crowd have specific needs
20 while others may just want to spend some time shopping for random things. There are
electronic surveillance devices which are provided at the entry of such market arenas, for
example, indoor arenas like malls, which record the number of people entering and leaving the
premises, but however this method doesn’t record the actual movement of the crowd inside the
mall. Furthermore, surveillance cameras are also positioned in different areas that record the
25 movement of the crowd, but only provides a manual visual idea of whether a certain group in
the crowd is concentrated in a certain area of the shopping mall, or whether an individual is
visiting a store periodically, etc. However, this fails to provide information regarding the actual
movement of the crowd inside the mall.
30 In the current scenario, the tracking of crowd movement in indoor arenas using location
tracking is an emerging technology that enables businesses such as retailers and hoteliers to
better track movement patterns of visitors. This enables business to better understand key
metrics, for example, in and out count, dwell time that denotes how much time is spent at
3
specific locations, as well as path navigation maps that generally indicate the most popular
routes taken by the user once inside the facility. These metrics in turn helps businesses to
optimize operations in many different ways. Maximizing customer safety, scheduling of
optimal workforce placements and product placements are just a few examples. Other examples
5 include the much-needed application in the field of public safety where a method or device is
required to prevent stampedes that may occur as a result of unmonitored movement of crowd
in a certain type of market place. Such stampede can only be monitored if there is a real time
analytical procedure to monitor the nature and density of crowd movement in such type of
market places.
10
However, there are certain problems faced during tracking of the movement of visitors in an
indoor environment. These problems include (1) Localization of visitor, or Visitor’s device, in
indoor environment, (2) Tracking movement of visitor from one location to another, (3)
Combining of individual movement patterns into group dynamics in the context of a specific
15 indoor layout, (4) the requirement of user privacy, since the current systems because do not
track individual visitors, but rather massed movements, etc. Therefore, there is a need for a
method and devices associated with the method that overcomes the aforementioned problems.
Therefore, there is a need to apply user dynamics through Wi-Fi device localization in an indoor
20 environment to monitor crowd movement within the indoor environment. Each problem that is
addressed in the conventional systems needs to be dealt separately. For the problem of visitor
localization, various approaches based on Cellular, Bluetooth, and Wi-Fi technologies are used
for indoor localization. Wi-Fi RSSI (Received Signal Strength Indicator) fingerprinting based
methods are popular because they do not require active participation from tracked devices and
25 require minimal infrastructure cost. However, these approaches are constrained by the
differences in the transmission characteristics of Wi-Fi devices. Differences in transmission
characteristics result in sub-optimal localization accuracy.
In addition to the matter above, for tracking the movement of Wi-Fi devices across various
30 locations, the preferred method is to track individual devices using a common identifier such
as the Wi-Fi MAC address. However, this infringes on the user’s privacy. Furthermore, modern
devices employ MAC randomization wherein instead of using the actual MAC of the device, a
randomly generated MAC address is used in Wi-Fi packets when the device is not connected
4
to any access point. Since the MAC address used by a device keeps changing, tracking the
journey of the device across different locations becomes very difficult. Hence, there is a need
for a method or a system for improving indoor localization and device tracking accuracy,
considering the differences in transmission characteristics of devices and MAC randomization
5 performed by modern phones.
SUMMARY OF THE INVENTION
It is intended that all such features, and advantages be included within this description, be
within the scope of the present invention, and be protected by the accompanying claims. The
10 following summary is provided to facilitate an understanding of some of the innovative features
unique to the disclosed embodiment and is not intended to be a full description. A full
appreciation of the various aspects of the embodiments disclosed herein can be gained by
taking the entire specification, claims, drawings, and abstract as a whole.
15 The indoor location positioning system disclosed here addresses the need for a method and a
system for improving indoor localization and device tracking accuracy, considering the
differences in transmission characteristics of devices and MAC randomization performed by
modern phones.
20 An indoor location positioning system comprises at least one processor, Wi-Fi scanners, and a
RSSI fingerprint database. The Wi-Fi scanners and the RSSI fingerprint database are
configured to work with the at least one processor. The Wi-Fi scanners are positioned at a
plurality of locations within an environment, wherein the environment is partitioned into
interconnected zones, wherein the Wi-Fi scanners receive anonymous probe messages from
25 Wi-Fi devices that are present in the environment during a calibration phase of the indoor
location positioning system. The RSSI values of the anonymous probe messages received by
the Wi-Fi scanners are collated to create RSSI fingerprints corresponding to the different
locations respectively. One or more of the Wi-Fi scanners are used as anchor scanners to
counter the difference in transmit power levels of different Wi-Fi devices that result in different
30 RSSI fingerprints for the different types of Wi-Fi devices, wherein the RSSI value at each
anchor scanner is subtracted from the RSSI values at rest of the Wi-Fi scanners to generate the
RSSI fingerprint.
5
The RSSI fingerprint database is generated after the calibration phase that contains the
generated RSSI fingerprints corresponding to the different locations. The processor, after the
calibration phase, matches the RSSI fingerprint of the anonymous probe messages from the
Wi-Fi devices against the RSSI fingerprints recorded during the calibration phase to predict
5 the location of the Wi-Fi device in a prediction phase. In an embodiment, the indoor location
positioning system further comprises a first correction procedure that involves correction for
Wi-Fi channel dependency. The Wi-Fi scanners that are used in the location prediction are
tuned to a same channel instead of performing frequency hopping to counter impact of
differences of impairment on different channels resulting in different RSSI values on the
10 different channels.
In an embodiment, the indoor location positioning system further comprises a second
correction that involves correction for Wi-Fi orientation of the Wi-F devices. Different
orientations of the Wi-Fi devices are used during the calibration phase to counter inconsistency
15 in accuracy during the prediction phase because the RSSI values change with change in the
orientation of the Wi-Fi device during the calibration phase and location prediction phase. A
third correction is also disclosed that involves the usage of the anchor scanners to counter the
difference in transmit power levels of different Wi-Fi devices. In an embodiment, the indoor
location positioning system further comprises a fourth correction that involves mapping user
20 data associated with each of the W-Fi devices to an indoor topology of the environment and
use the mapping to measure group dynamics of the Wi-Fi devices.
In view of the above, an independent method is disclosed for indoor location positioning of
Wi-Fi devices using RSSI fingerprinting of anonymous probe messages transmitted by Wi-Fi
25 devices. RSSI fingerprinting methods require training/calibration on each new place since no
two indoor environments have the same signal propagation characteristics. Based on the
disclosed method, the accuracy of location prediction does not change if the type of training
device is different from the type of tracked device. The method comprises the following steps:
i. Creating RSSI fingerprint by collecting anonymous probe messages transmitted by Wi30 Fi device on strategically placed scanners tuned to one single Wi-Fi channel
ii. Collecting RSSI fingerprint by placing training Wi-Fi device in different orientations
at the same location.
6
iii. Use of differential RSSI in creating RSSI fingerprint instead of using actual RSSI value
received at each scanner.
Subsequently demonstrating the use of the above method in measuring bulk movement of users
5 in an indoor arena with a known topology. By taking the predicted location as an outcome of
the RSSI fingerprinting activity and mapping it to a known indoor topology, we can track group
dynamics of users, including arrival and departure rates, the ‘stability’ of the indoor population,
concentration hotspots and the flow of users from one point to another within the indoor area
in a secure and anonymous manner. By adding more and more monitors, we can get more
10 accuracy using the same basic set of tools.
BRIEF DESCRIPTION OF DRAWINGS
The invention can be better understood with reference to the following drawings. The
components in the drawings are not necessarily to scale, emphasis instead being placed upon
15 clearly illustrating the principles of the present invention. Moreover, in the drawings, like
reference numerals designate corresponding parts throughout the several views.
Figure 1A is the schematic view of Wi-Fi Scanner placement in an indoor arena, which further
describes the working of the Wi-Fi RSSI fingerprinting technique.
20
Figure 1B is a system diagram corresponding to the the schematic view of Wi-Fi Scanner
placement in an indoor arena as shown in the Figure 1A, which also describes the working of
the Wi-Fi RSSI fingerprinting technique.
25 Figure 2 is a converted stylized graph of the schematic view of the layout of the indoor arena
as shown in Figure 1A.
Figure 3 is a graphical view showing the de-spread occupancy data for all edges in an indoor
topology that is shown in Figure 2.
30
Figure 4 is a graphical view showing correlation between edges, and further showing relative
drift rate over the network based on the graphical view in Figure 2.
7
Figure 5 is a method flow diagram that explains the step-by-step process associated with indoor
location positioning using Wi-Fi scanners in the Wi-Fi RSSI fingerprinting technique.
DESCRIPTION OF THE INVENTION
5
The aim of the present invention is to track crowd movement in indoor arenas using location
tracking, wherein user dynamics is applied through Wi-Fi device localization in an indoor
environment. The foregoing advantages as well as the working of the method that involves the
application of user dynamics for Wi-Fi device localization and components associated with the
10 method will become more noticeable and understandable from the following detail description
thereof when read in conjunction with the accompanying drawings.
Regarding Figures 1A-1B and Figure 5, Figure 1A is the schematic view of Wi-Fi Scanner 102
placement in an indoor arena 100, which further describes the working of the Wi-Fi RSSI
15 fingerprinting technique. Figure 1B is a system diagram corresponding to the schematic view
of Wi-Fi scanners 102 that are positioned in an indoor arena 100 as shown in the Figure 1A
and described in Figure 5, which also describes the working of the Wi-Fi RSSI fingerprinting
technique. The Wi-Fi RSSI fingerprinting technique used here resolves the problem of Wi-Fi
device localization, for example, a mobile phone. This method includes learning RSSI
20 fingerprints or RSSI patterns for areas/zones of interest in any indoor environment 100 and
consequently using the learnt association between RSSI values and position data to make
precise predictions of device locations in unmarked readings. The Wi-Fi RSSI fingerprinting
technique consists of two phases: (a) calibration/training phase and (b) a location prediction
phase. In the method disclosed herein, the local area or the indoor environment 100 is
25 partitioned 506 into a graph-pattern of interconnected zones 101a, 101b, and 101c, where each
zone characterizes a coverage zone for the Wi-Fi access, as further described in Figure 2.
During the training phase, probes are generated that are associated with their location and the
corresponding fingerprints consisting of RSSI readings.
30 Furthermore, at least one processor 106 is provided 502 as shown in Figure 5, wherein the WiFi scanners 102 and an RSSI fingerprint database 108 are configured to work with the at least
one processor 106, as shown in Figure 1B. As explained above, the Wi-Fi scanners 102 are
positioned at multiple locations within the environment 100. The RSSI fingerprints are
8
configured to train machine learning based classification algorithms. Later on, the trained
classification algorithms are configured to predict the location of a Wi-Fi device 104 using
RSSI fingerprint. These RSSI fingerprinting methods necessitate training/calibration on each
new place since no two indoor environments 100 have the same signal propagation
5 characteristics. During the calibration phase, the Wi-Fi scanners 102 are placed at different
locations and are set up to record the RSSI value of received 508 anonymous probe messages.
The Wi-Fi scanners 102 are, for example, 102a, 102b, 102c, and 102d, as shown in Figure 1B.
In another embodiment, Wi-Fi devices 104a, 104b, 104c, and 104d, for example, are
transmitters that are placed at various locations of interest. RSSI values of anonymous probe
10 messages received by Wi-Fi scanners 102 are collated 510 to create the RSSI fingerprint
corresponding to a location. For example if n Wi-Fi scanners 102 are being used, and RSSI
value of a probe request is R_j at scanner S_j, where j∈[1..n] a typical RSSI fingerprint tuple
with RSSI values from all n Wi-Fi scanners 102 would be
S_1:R_1,S_2:R_2,S_3:R_3,S_4:R_4,S_5:R_5,S_6:R_6,.......S_n:R_n. The output of the
15 calibration/training phase is an RSSI fingerprint database 108 corresponding to each location
of interest is as shown in Figures 1A-1B. In the location prediction phase, the RSSI fingerprint
of probe messages from a Wi-Fi device 104 is used to perform a match against the RSSI
fingerprint recorded during the calibration phase. Based on this match, the location of the
transmitting Wi-Fi device 104 is predicted.
20
The basic premise of RSSI fingerprinting method is that the combination of RSSI values
measured by the different Access Points are unique for each learnt location and remains
constant across different categories of Wi-Fi devices 104a, 104b, 104c, and 104d placed at that
location and other environmental changes. The Wi-Fi devices 104a, 104b, 104c, and 104d are,
25 for example, Wi-Fi device 1, 2, 3, and 4 as shown in Figure 1B. However, tests advise that
apart from distance between the transmitting Wi-Fi device 104 and Wi-Fi scanner 102 and
propagation characteristics of indoor environment 100, the RSSI fingerprint changes with: (1)
Wi-Fi channel being used by transmitting device and Wi-Fi scanner 102, (2) Orientation of WiFi device 104, and (3) Wi-Fi chipset used in device. If any of the above aspects during location
30 prediction phase is different from those during the calibration phase, the accuracy of location
prediction is considerably compromised. Change in RSSI fingerprint of mobile devices using
different chipsets propose that the calibration phase should include all probable types of devices
9
that are used during location prediction. For any deployment scenario, it is not possible to use
every type of device during the calibration phase.
Considering the above-mentioned problems faced in the RSSI fingerprint techniques, there are
5 certain enhancements that are provided to the RSSI fingerprinting process to overcome the
problems. A set of corrections are applied to improvise indoor positioning, and by using these
corrections, the location prediction accuracy of the system is significantly improved. The first
correction involves correction for Wi-Fi channel dependency. Here, owing to different path
loss characteristics of different Wi-Fi channels, RSSI values of Wi-Fi packets on different
10 channels differ significantly. To counter 512 the impact of the difference in RSSI values on
different channels, all the Wi-Fi scanners 102a, 102b, 102c, and 102d that are used in location
tracking is tuned to the same channel instead of performing frequency hopping.
The second correction involves correction for Wi-Fi orientation, where the RSSI value changes
15 with change in the orientation of the Wi-Fi device 104, for example, a mobile phone. This
impacts the accuracy of prediction. If the phone orientation during the calibration phase and
location prediction phase is different, the prediction accuracy is usually not high. To mitigate
this issue, different phone orientations are used during the calibration phase. The third
correction involves correction for Wi-Fi device 104 type, where each Wi-Fi device or chipset,
20 has different transmission characteristics. These differences result in different RSSI
fingerprints for different types of devices 104, even when they are placed at the same location
in the same orientation. To overcome the issues caused by the difference in transmit power
level of devices 104, one of the Wi-Fi scanners 102 is used as an anchor scanner, for example,
anchor scanner 102b as shown in Figure 1B.
25
For generating an RSSI fingerprint, instead of using absolute RSSI value at every scanner 102,
the RSSI value at the anchor scanner 102b is subtracted 514 from RSSI value at every other
scanner 102a, 102c, or 102d. The fourth and the final correction involves to map the user data
to an indoor topology and use it to measure group dynamics. This involves three sub-steps,
30 which is described in Figures 2, 3, 4, and 5. In view of the above, the RSSI fingerprint database
108 is generated 516 after the calibration phase that contains the generated RSSI fingerprints
corresponding to the different locations. Thereafter, the processor 106 matches 518 the RSSI
fingerprint of the anonymous probe messages from the Wi-Fi devices 104 against the RSSI
10
fingerprints recorded during the calibration phase to predict the location of the Wi-Fi device
104 in a prediction phase.
Figure 2 is a converted stylized graph of the schematic view of the layout of the indoor arena
5 100 as shown in Figures 1A-1B. RSSI fingerprinting process is used to convert each received
probe, as described in Figures 1A-1B, to a likely location on the graph that is, where location
refers to one of the edges S1, S2, etc. The group dynamic information is also measured in terms
of the graph for instance, the current occupancy of the edge S1, or the transition probability of
S6 to S7.
10
Figure 3 is a graphical view showing the de-spread occupancy data for all edges in an indoor
topology that is shown in Figure 2. The second step involves de-spreading of the received probe
data, using a raised cosine filter, so as to generate an estimate of the occupancy of each edge
as a continuous function Si(𝑡) for each ith edge.
15
Figure 4 is a graphical view showing correlation between edges S1, S2, etc., and further
showing relative drift rate over the network based on the graphical view in Figure 2. The final
stage involves the estimation of group dynamics by tracking the occupancy of each edge
against the adjacent edges, using standard time-series techniques. For example, Figure 4 shows
20 the correlation between adjacent edges S1 and S2 and then between S2 and S5 (as shown in
the graph of Figure 2). By comparing the estimation peaks, a user measures bulk statistic, such
as the drift rate between the two edges.
As will be appreciated by one of skill in the art, the present invention may be embodied as a
25 method, system and apparatus. Accordingly, the present invention may take the form of an
entirely hardware embodiment, a software embodiment or an embodiment combining software
and hardware aspects.
It will be understood that each block of the block diagrams, can be implemented by computer
30 program instructions. These computer program instructions may be provided to a processor of
a general-purpose computer, special purpose computer, or other programmable data processing
apparatus to produce a machine, such that the instructions, which execute via the processor of
11
the computer or other programmable data processing apparatus, create means for implementing
the functions/acts specified in the flowchart and/or block diagram block or blocks.
In the drawings and specification, there have been disclosed exemplary embodiments of the
5 invention. Although specific terms are employed, they are used in a generic and descriptive
sense only and not for purposes of limitation of the scope of the invention.

We Claim:

1. An indoor location positioning system comprising:
at least one processor, wherein a set of Wi-Fi scanners and an RSSI fingerprint database
are configured to work with the at least one processor;
the Wi-Fi scanners are positioned at a plurality of locations within an environment,
wherein the environment is partitioned into interconnected zones, wherein the Wi-Fi scanners
receive anonymous probe messages from Wi-Fi devices that are present in the environment
during a calibration phase of the indoor location positioning system, and wherein RSSI values
of the anonymous probe messages received by the Wi-Fi scanners are collated to create RSSI
fingerprints corresponding to the different locations respectively;
wherein one or more of the Wi-Fi scanners are used as anchor scanners to counter the
difference in transmit power levels of different Wi-Fi devices that result in different RSSI
fingerprints for the different types of Wi-Fi devices, wherein the RSSI value at each anchor
scanner is subtracted from the RSSI values at rest of the Wi-Fi scanners to generate the RSSI
fingerprint;
the RSSI fingerprint database is generated after the calibration phase that contains the
generated RSSI fingerprints corresponding to the different locations; and
the processor, after the calibration phase, matches the RSSI fingerprint of the
anonymous probe messages from the Wi-Fi devices against the RSSI fingerprints recorded
during the calibration phase to predict the location of the Wi-Fi device in a prediction phase.
2. The indoor location positioning system claimed in claim 1, further comprising a first
correction procedure that involves correction for Wi-Fi channel dependency, wherein the WiFi scanners that are used in the location prediction are tuned to a same channel instead of
performing frequency hopping to counter impact of differences of impairment on different
channels resulting in different RSSI values on the different channels.
3. The indoor location positioning system claimed in claim 1, further comprising a second
correction that involves correction for Wi-Fi orientation of the Wi-F devices, wherein
different orientations of the Wi-Fi devices are used during the calibration phase to counter
inconsistency in accuracy during the prediction phase because the RSSI values change with
change in the orientation of the Wi-Fi device during the calibration phase and location
13
prediction phase, and followed by a third correction that involves the usage of the anchor
scanners to counter the difference in transmit power levels of different Wi-Fi devices.
4. The indoor location positioning system claimed in claim 1, further comprising a fourth
correction that involves mapping user data associated with each of the W-Fi devices to an
indoor topology of the environment, and use the mapping to measure group dynamics of the
Wi-Fi devices.
5. A method for indoor location positioning comprising:
providing at least one processor, wherein a set of Wi-Fi scanners and an RSSI
fingerprint database are configured to work with the at least one processor;
positioning Wi-Fi scanners at a plurality of locations within an environment;
partitioning the environment into a into interconnected zones;
receiving, via the Wi-Fi scanners, anonymous probe messages from Wi-Fi devices that
are present in the environment during a calibration phase;
collating RSSI values of the anonymous probe messages received by the Wi-Fi
scanners to create RSSI fingerprints corresponding to the different locations respectively;
countering difference in transmit power levels of different Wi-Fi devices by using one
or more of the Wi-Fi scanners as anchor scanners, which result in different RSSI fingerprints
for the different types of Wi-Fi devices;
subtracting the RSSI value at each anchor scanner from the RSSI values at rest of the
Wi-Fi scanners to generate the RSSI fingerprint;
generating the RSSI fingerprint database after the calibration phase that contains the
generated RSSI fingerprints corresponding to the different locations; and
matching, via the processor after the calibration phase, the RSSI fingerprint of the
anonymous probe messages from the Wi-Fi devices against the RSSI fingerprints recorded
during the calibration phase to predict the location of the Wi-Fi device in a prediction phase.
6. The method as claimed in claim 5, further comprising correcting for Wi-Fi channel
dependency via a first correction procedure, wherein the Wi-Fi scanners that are used in the
location prediction are tuned to a same channel instead of performing frequency hopping to
counter impact of differences of impairment on different channels resulting in different RSSI
values on the different channels.
14
7. The method as claimed in claim 5, further comprising correcting for Wi-Fi orientation of
the Wi-F devices via a second correction, wherein different orientations of the Wi-Fi devices
are used during the calibration phase to counter inconsistency in accuracy during the
prediction phase because the RSSI values change with change in the orientation of the Wi-Fi
device during the calibration phase and location prediction phase, , and followed by a third
correction that involves the usage of the anchor scanners to counter the difference in transmit
power levels of different Wi-Fi devices.
8. The method as claimed in claim 5, further comprising mapping user data associated with
each of the W-Fi devices via a fourth correction, to an indoor topology of the environment,
and use the mapping to measure group dynamics of the Wi-Fi devices.