U.S. patent application number 15/084799 was filed with the patent office on 2017-10-05 for autonomous semantic labeling of physical locations.
This patent application is currently assigned to Intel Corporation. The applicant listed for this patent is Intel Corporation. Invention is credited to JONATHAN J. HUANG, HONG LU, LAMA NACHMAN, RAHUL C. SHAH, CHIEH-YIH WAN, LEI YANG.
Application Number | 20170289768 15/084799 |
Document ID | / |
Family ID | 59928453 |
Filed Date | 2017-10-05 |
United States Patent
Application |
20170289768 |
Kind Code |
A1 |
WAN; CHIEH-YIH ; et
al. |
October 5, 2017 |
AUTONOMOUS SEMANTIC LABELING OF PHYSICAL LOCATIONS
Abstract
A portable electronic device may generate a (RF) radio frequency
fingerprint that includes information representative of at least a
portion of RF signals received at a given physical location. The RF
fingerprint may include, for example, a unique identifier and a
signal strength that are both logically associated with at least a
portion of the received RF signals. The portable electronic device
may also receive data representative of a number of environmental
parameters about the portable electronic device. These
environmental parameters may be measured using sensors carried by
the portable electronic device. Considered in combination, these
environmental parameters provide an environmental signature for a
given location. When combined into a data cluster, the RF
fingerprint and the environmental signature may provide an
indication of the physical subdivision where the portable
electronic device is located. The portable electronic device may
then generate a proposed semantic label for the physical
subdivision.
Inventors: |
WAN; CHIEH-YIH; (Beaverton,
OR) ; YANG; LEI; (Hillsboro, OR) ; HUANG;
JONATHAN J.; (Pleasanton, CA) ; LU; HONG;
(Santa Clara, CA) ; SHAH; RAHUL C.; (San
Francisco, CA) ; NACHMAN; LAMA; (Santa Clara,
CA) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
Intel Corporation |
Santa Clara |
CA |
US |
|
|
Assignee: |
Intel Corporation
Santa Clara
CA
|
Family ID: |
59928453 |
Appl. No.: |
15/084799 |
Filed: |
March 30, 2016 |
Current U.S.
Class: |
1/1 |
Current CPC
Class: |
H04W 4/023 20130101;
H04W 4/029 20180201; H04W 4/33 20180201; H04W 4/70 20180201; H04W
4/38 20180201; H04W 4/80 20180201; H04L 67/303 20130101; H04W 4/30
20180201 |
International
Class: |
H04W 4/04 20060101
H04W004/04; H04L 29/08 20060101 H04L029/08; H04W 4/02 20060101
H04W004/02; H04W 4/00 20060101 H04W004/00 |
Claims
1. A semantic labeling controller, comprising: a transceiver to
receive a plurality of radio frequency (RF) signals within a
defined portion of the RF spectrum; a sensor to generate a signal
that includes data or information indicative of an ambient
environment in a physical subdivision external to the controller;
RF fingerprint generator circuitry coupled to the transceiver, the
RF fingerprint generator circuitry to receive a plurality of RF
signals and selectively generate an RF fingerprint that includes
data indicative of at least one parameter associated with each of
the received RF signals; environmental signature generator
circuitry coupled to the sensor, the environmental signature
generator circuitry to receive the signal that includes data or
information indicative of the ambient environment in the physical
subdivision and selectively generate an environmental signature
that includes data indicative of the ambient environment external
to the controller; data grouping circuitry coupled to the RF
fingerprint generator circuitry and to the environmental signal
generator circuitry to selectively group the RF fingerprint and the
environmental signature into a data cluster responsive to the RF
fingerprint meeting at least one threshold condition; and semantic
labeling circuitry coupled to the data grouping circuitry to
autonomously generate a proposed semantic label for logical
association with the physical subdivision based at least in part on
the RF fingerprint and the environmental signature included in the
data cluster.
2. The semantic labeling controller of claim 1 wherein the
transceiver, in operation, receives RF signals within at least a
portion of an RF spectrum defined by Bluetooth.RTM. or Near Field
Communication (NFC) standards.
3. The semantic labeling controller of claim 1 wherein the
transceiver, in operation, receives RF signals within at least a
portion of an RF spectrum defined by IEEE 802.11 (WiFi)
standards.
4. The semantic labeling controller of claim 3 wherein the RF
fingerprint generator circuitry selectively generates the RF
fingerprint using at least a unique identifier and a signal
strength logically associated with each respective one of the
plurality of received RF signals.
5. The semantic labeling controller of claim 4 wherein the semantic
labeling circuitry further: generates a human perceptible output
that includes data indicative of the proposed semantic label;
receives input indicative of a confirmation of the proposed
semantic label; and responsive to receipt of the input indicative
of the confirmation of the proposed semantic label, logically
associates the confirmed semantic label with the physical
subdivision.
6. The semantic labeling controller of claim 5 wherein the sensor
comprises at least one of: an audio input sensor, an ambient light
sensor, an accelerometer, or an image sensor.
7. The semantic labeling controller of claim 6, further comprising
device profile selection circuitry coupled to the semantic labeling
circuitry to select a device profile based at least in part on the
identified physical subdivision.
8. The semantic labeling controller of claim 6, further comprising
device detection circuitry coupled to the semantic labeling
circuitry and to the data grouping circuitry, the device detection
circuitry to: detect a presence of at least one other local
portable electronic device; receive a signal from the at least one
other local portable electronic device, the signal including data
representative of a semantic label and an identified data cluster
logically associated with the semantic label; compare the received
data representative of the identified data cluster with the data
cluster generated by the data grouping circuitry; and responsive to
a favorable comparison, logically associate the received semantic
label with the physical subdivision.
9. The semantic labeling controller of claim 1 wherein the data
grouping circuitry selectively determines an RF signal sampling
rate based at least in part on at least one of data or information
included in the output signal from the sensor.
10. The semantic labeling controller of claim 9 wherein the data
grouping circuitry selectively reduces the RF signal sampling rate
responsive to data or information included in the output signal of
the sensor indicative of the sensor being stationary.
11. The semantic labeling controller of claim 9 wherein the data
grouping circuitry further selectively halts RF signal sampling
responsive to data or information included in the output signal of
the sensor indicative of the sensor being in motion.
12. The semantic labeling controller of claim 1 wherein the data
grouping circuitry selectively groups the RF fingerprint and the
environmental signature into a data cluster responsive to the RF
fingerprint remaining within an RF fingerprint threshold for
greater than a defined temporal threshold period.
13. A semantic labeling method, comprising: selectively generating,
by radio frequency (RF) fingerprint generator circuitry, an RF
fingerprint that includes data indicative of at least one parameter
associated with each of the received RF signals; generating, by a
sensor, a signal that includes data or information indicative of an
ambient environment in a physical subdivision external to the
controller; generating, by environmental signature generating
circuitry coupled to the sensor, an environmental signature that
includes data indicative of the ambient environment external to the
sensor; selectively grouping, by data grouping circuitry coupled to
the RF fingerprint generator circuitry and to the environmental
signal generator circuitry, the RF fingerprint and the
environmental signature into a data cluster responsive to the RF
fingerprint meeting at least one threshold condition; and
autonomously generating, by semantic labeling circuitry coupled to
the data grouping circuitry, a proposed semantic label for logical
association with the physical subdivision based at least in part on
the RF fingerprint and the environmental signature included in the
data cluster.
14. The method of claim 13 wherein selectively generating an RF
fingerprint that includes data indicative of at least one parameter
associated with each of the received RF signals comprises:
selectively generating, by the RF fingerprint generator circuitry,
an RF fingerprint using at least one parameter from each of the
plurality of received RF signals, each of the plurality of received
RF signals within at least a portion of an RF spectrum that
includes at least one of: a Bluetooth.RTM. RF spectrum or Near
Field Communication (NFC) RF spectrum.
15. The method of claim 13 wherein selectively generating an RF
fingerprint that includes data indicative of at least one parameter
associated with each of the received RF signals comprises:
selectively generating, by the RF fingerprint generator circuitry,
an RF fingerprint using at least one parameter from each of the
plurality of received RF signals, each of the plurality of received
RF signals within at least a portion of an RF spectrum that
includes an IEEE 802.11 (WiFi) RF spectrum.
16. The method of claim 15 wherein selectively generating at least
one RF fingerprint using at least one parameter from each of the
plurality of received RF signals comprises: selectively generating,
by the RF fingerprint generator circuitry, at least one RF
fingerprint using at least a unique identifier and a signal
strength logically associated with each respective one of the
plurality of received RF signals.
17. The method of claim 13 wherein autonomously generating a
proposed semantic label for logical association with the physical
subdivision based at least in part on the RF fingerprint and the
environmental signature included in the data cluster comprises:
generating, by the semantic labeling circuitry, a human perceptible
output that includes data indicative of the proposed semantic
label; receiving, by the semantic labeling circuitry, an input
indicative of a confirmation of the proposed semantic label; and
logically associating, by the semantic labeling circuitry, the
confirmed semantic label with the physical subdivision responsive
to receiving the input indicative of the confirmation of the
semantic label.
18. The method of claim 13 wherein generating an environmental
signature that includes data indicative of the ambient environment
external to the sensor comprises: receiving, by a sensor, an
environmental signature that includes data or information
indicative of at least one of: an audible event external to the
portable electronic device, an ambient light event external to the
portable electronic device, an ambient motion event external to the
portable electronic device, or an image of the environment external
to the portable electronic device.
19. The method of claim 13, further comprising: selecting, by
device profile selection circuitry coupled to the semantic labeling
circuitry, a portable electronic device profile based at least in
part on the identified physical subdivision.
20. The method of claim 13, further comprising: detecting, by
device detection circuitry coupled to the semantic labeling
circuitry and to the data grouping circuitry, a presence of at
least one other local portable electronic device; receiving, by the
device detection circuitry, a signal from the at least one other
local portable electronic device, the signal including data
representative of a semantic label and an identified data cluster
logically associated with the semantic label; comparing, by the
device detection circuitry, the received data representative of the
identified data cluster with the data cluster generated by the data
grouping circuitry; and responsive to a favorable comparison,
logically associating, by the device detection circuitry, the
received semantic label with the physical subdivision.
21. The method of claim 13 wherein selectively generating at least
one RF fingerprint using at least one parameter from each of the
plurality of received RF signals comprises: selectively
determining, by the data grouping circuitry, an RF signal sampling
rate based at least in part on at least one of data or information
included in an output signal from the sensor.
22. The method of claim 21 wherein selectively determining an RF
signal sampling rate based at least in part on at least one of data
or information included in an output signal from the sensor
comprises: selectively reducing the RF signal sampling rate, by the
data grouping circuitry, responsive to data or information included
in the output signal of the sensor indicative of the sensor being
stationary.
23. The method of claim 21 wherein selectively determining an RF
signal sampling rate based at least in part on at least one of data
or information included in an output signal from the sensor
comprises: selectively halting RF signal sampling, by the data
grouping circuitry, responsive to data or information included in
the output signal of the sensor indicative of the sensor being in
motion.
24. A storage device that includes machine-readable instructions
that, when executed by a circuit, cause the circuit to provide a
semantic labeling controller, that: selectively generates an RF
fingerprint that includes data indicative of at least one parameter
associated with each of the received RF signals; generates a signal
that includes data or information indicative of an ambient
environment in a physical subdivision external to the controller;
generates an environmental signature that includes data indicative
of the ambient environment external to the sensor; selectively
groups, the RF fingerprint and the environmental signature into a
data cluster responsive to the RF fingerprint meeting at least one
threshold condition; and autonomously generates a proposed semantic
label for logical association with the physical subdivision based
at least in part on the RF fingerprint and the environmental
signature included in the data cluster.
25. The method of claim 24, further comprising machine-readable
instructions that cause the semantic labeling controller to:
detecting a presence of at least one other local portable
electronic device; receiving a signal from the at least one other
local portable electronic device, the signal including data
representative of a semantic label and an identified data cluster
logically associated with the semantic label; comparing the
received data representative of the identified data cluster with
the data cluster generated by the data grouping circuitry; and
responsive to a favorable comparison, logically associating the
received semantic label with the physical subdivision.
Description
TECHNICAL FIELD
[0001] The present disclosure relates to autonomous location
detection for portable devices.
BACKGROUND
[0002] Currently, many portable electronic devices use absolute
location or geocoordinates determined or otherwise calculated using
traditional IEEE 802.11 (WiFi.RTM.) location services. These
techniques usually derive device location using trilateration based
on a received signal strength indicator (RSSI) propagation model, a
WiFi time-of-flight (ToF), or by compiling a set of RSSIs from a
number of different WiFi sources as a location signature (e.g.,
RSSI fingerprinting). Some of the solutions permit the user of the
portable electronic device to enter or otherwise input semantic
labels (e.g., indoor locations at a room level) that are then
logically associated with a respective location.
[0003] A typical approach may include collecting WiFi samples at a
number of locations within each room or semantic location followed
by applying a supervised classification algorithm to map a new WiFi
measurement to the trained hotspots. Such existing solutions are
tedious, requiring the portable electronic device user to manually
train a room or semantic classifier to create a WiFi database that
maps a particular signature to a particular semantic or room
location. Such a system is labor-intensive and requires periodic
maintenance when the room environment changes.
BRIEF DESCRIPTION OF THE DRAWINGS
[0004] Features and advantages of various embodiments of the
claimed subject matter will become apparent as the following
Detailed Description proceeds, and upon reference to the Drawings,
wherein like numerals designate like parts, and in which:
[0005] FIG. 1 provides a schematic of an illustrative autonomous
semantic labeling system, in accordance with at least one
embodiment of the present disclosure;
[0006] FIG. 2 provides a schematic of an illustrative autonomous
semantic labeling system using multiple devices in a room, in
accordance with at least one embodiment of the present
disclosure;
[0007] FIG. 3 provides a block diagram of an example system on
which the autonomous semantic labeling system may be implemented,
in accordance with at least one embodiment of the present
disclosure;
[0008] FIG. 4 provides a high-level block flow diagram of an
example semantic labeling method, in accordance with at least one
embodiment of the present disclosure;
[0009] FIG. 5 provides a high-level block flow diagram of an
example environmental signature content extraction system for use
with the semantic labeling method in FIG. 4, in accordance with at
least one embodiment of the present disclosure;
[0010] FIG. 6 provides a high-level logic flow diagram of an
example semantic labeling method, in accordance with at least one
embodiment of the present disclosure;
[0011] FIG. 7 provides a high-level logic flow diagram of an
example semantic labeling method that includes autonomous device
profile selection, in accordance with at least one embodiment of
the present disclosure;
[0012] FIG. 8 provides a high-level logic flow diagram of an
example semantic labeling method that includes exchanging physical
location data with other local portable electronic devices, in
accordance with at least one embodiment of the present disclosure;
and
[0013] FIG. 9 provides a high-level logic flow diagram of an
example semantic labeling method in which the radio frequency (RF)
sampling frequency is adjusted based on a physical displacement of
the portable electronic device, in accordance with at least one
embodiment of the present disclosure.
[0014] Although the following Detailed Description will proceed
with reference being made to illustrative embodiments, many
alternatives, modifications and variations thereof will be apparent
to those skilled in the art.
DETAILED DESCRIPTION
[0015] The systems and methods disclosed herein provide solutions
that automatically learn the semantic labels that are logically
associated with physical locations in an indoor environment. Such
physical locations may include rooms or similar divisible
sub-structures existing within a larger structure or may include
sub-portions of a larger area such as a sub portion of gymnasium or
similar large, open, structure having few, if any, internal
dividers. The systems and methods disclosed herein propose semantic
labeling for physical locations without requiring any explicit user
training. Such systems and methods beneficially and advantageously
provide a smooth user experience that incorporates location aware
technology on a day-to-day basis.
[0016] Such semantic labeling of physical locations offers
significant benefits over current state-of-the-art systems. For
example, such systems and methods may provide parental control in a
home environment where a portable electronic device may
automatically switch between device profiles simply by recognizing
a physical location such as the master bedroom, living room, or
children's bedroom. In another example, such systems and methods
may be useful for sleep monitoring by tracking locations where the
device user sleeps and providing feedback on the quality of the
user's sleep by location. In another example, such systems and
methods may be used for elder care by tracking where elders spend
time to develop an understanding of the elder's daily routine such
as meals, bathroom activity, and entertainment. Such may permit the
detection of unhealthy trends and emergencies.
[0017] Instead of requiring the user of the portable electronic
device to manually collect and label WiFi training data in each
room, the systems and methods described herein leverage WiFi data
collected in the background during the day-to-day activity of the
user. The systems and methods described herein automatically group
or cluster the WiFi signals into virtual rooms in a WiFi signal
space that may be mapped into physical rooms via a simplified user
annotation process. The systems and methods described herein may
employ data collected from one or more sensors carried by the
portable electronic device to infer user activity and filter data
based on a motion state of a user to reduce WiFi signal noise and
accommodate routine fluctuations in WiFi signals. Additionally,
such user mobility data may be used to adjust the sampling
frequency of WiFi data, identify personally significant places by
user, and determine optimal times to opportunistically collect
device user feedback. For example, the WiFi sampling frequency may
be reduced when device sensor data indicates the user is
stationary--such may reduce the collection of redundant data at a
single location. In another example, WiFi data collected when
device sensor data indicates the user is in motion may be discarded
since such data is typically collected when the user is in
transition between rooms. In embodiments, the filtered WiFi data
may be buffered in the portable electronic device memory until the
data size exceeds a defined threshold. When the collected data
exceeds the threshold value, an unsupervised clustering process may
be applied to the collected WiFi data to form logical clusters and
virtual rooms.
[0018] Once the user of the portable electronic device is
determined to be stationary (i.e., remaining in the same location
for greater than a defined time), the systems and methods described
herein would opportunistically and autonomously prompt the user to
provide feedback. Such feedback may be used to label the current
virtual room with a semantic label, confirm the system has
autonomously assigned a correct semantic label to the current
virtual room, or to correct and retrain the system when the system
has autonomously assigned an incorrect semantic label to the
current virtual room. Such device user feedback may beneficially be
used to training feedback for the classification state machine as
well as ground truth to evaluate the performance of the
classification state machine. A high performance state machine may
be used to automatically merge newly found data clusters or virtual
rooms with known or previously labeled physical rooms in the
absence of user feedback. For example, if a WiFi fingerprint in a
virtual room may be classified with a high confidence as a known
physical room, then the virtual room can share the room label and
be mapped to the corresponding physical room.
[0019] The systems and methods beneficially fuses data from sensors
frequently carried by portable electronic devices (e.g., audio
sensors, accelerometers) to augment the WiFi data in capturing the
context representation of a room and leverage that to automatically
propagate labels or merge rooms. For example, an ambient audio
profile of a room is likely to be unique across different rooms and
may be used as a signature to identify physical rooms. Newly found
virtual rooms that share the same audio signature with a physical
room may then be merged or labeled automatically without user
input. Similarly, fusing accelerometer data, different virtual
rooms detected during the period of time when the portable
electronic device is stationary may be merged automatically, or
even labeled accordingly if the stationary time window overlaps
with the detection of a physical room with high confidence.
[0020] A portable electronic device is provided. The portable
electronic device may include a transceiver that, in operation,
receives a plurality of radio frequency (RF) signals within a
defined portion of the RF spectrum, a sensor, and a circuit
communicably coupled to the transceiver and to the sensor. The
portable electronic device further includes a storage device
communicably coupled to the circuit, the storage device to store
machine readable instructions that, when executed by the at least
one circuit, cause the at least one circuit to provide a semantic
labeling controller. The semantic labeling controller selectively
generates at least one RF fingerprint using at least one parameter
from each of the plurality of received RF signals; receives from
the sensor at least one output signal that includes data or
information indicative of a physical subdivision external to the
portable electronic device; selectively generates at least one
environmental signature using at least one parameter from the at
least one sensor output signal; selectively groups data
representative of the at least one RF fingerprint and the data or
information representative of the at least one environmental
signature into a data cluster responsive to the RF fingerprint
meeting at least one threshold criteria; and autonomously proposes
a semantic label for logical association with a physical
subdivision based at least in part on the RF fingerprint and the
environmental signature.
[0021] A portable electronic device controller is also provided.
The controller may include a circuit communicably coupleable to a
radio frequency (RF) transceiver and to a sensor that generates a
signal that includes data or information indicative of an ambient
environment in a physical subdivision about the circuit and a
storage device communicably coupled to the circuit, the storage
device to store machine readable instructions that, when executed
by the at least one circuit, cause the at least one circuit to
provide a semantic labeling controller. The semantic labeling
controller may selectively generate at least one RF fingerprint
using at least one parameter from each of the plurality of received
RF signals; receive from the sensor at least one output signal that
includes data or information indicative of a physical subdivision
external to the portable electronic device; selectively generate at
least one environmental signature using at least one parameter from
the at least one sensor output signal; selectively group data
representative of the at least one RF fingerprint and the data or
information representative of the at least one environmental
signature into a data cluster responsive to the RF fingerprint
meeting at least one threshold criteria; and autonomously proposes
a semantic label for logical association with a physical
subdivision based at least in part on the RF fingerprint and the
environmental signature.
[0022] An autonomous semantic labeling method is provided. The
method may include selectively generating, by a semantic labeling
controller, at least one radio frequency (RF) fingerprint using at
least one parameter from each of the plurality of received RF
signals; receiving, by the semantic labeling controller from a
communicably coupled sensor, at least one output signal that
includes data or information indicative of a physical subdivision
external to a portable electronic device; selectively generating,
by the semantic labeling controller, at least one environmental
signature using at least one parameter from the at least one sensor
output signal; selectively grouping, by the semantic labeling
controller, data representative of the at least one RF fingerprint
and the data or information representative of the at least one
environmental signature into a data cluster responsive to the RF
fingerprint meeting at least one threshold criteria; and
autonomously proposing, by the semantic labeling controller, a
semantic label for logical association with a physical subdivision
based at least in part on the RF fingerprint and the environmental
signature.
[0023] An autonomous semantic labeling system is provided. The
system may include a means for selectively generating at least one
radio frequency (RF) fingerprint using at least one parameter from
each of the plurality of received RF signals; a means for receiving
at least one output signal that includes data or information
indicative of a physical subdivision external to a portable
electronic device; a means for selectively generating at least one
environmental signature using at least one parameter from the at
least one sensor output signal; a means for selectively grouping
data representative of the at least one RF fingerprint and the data
or information representative of the at least one environmental
signature into a data cluster responsive to the RF fingerprint
meeting at least one threshold criteria; and a means for
autonomously proposing a semantic label for logical association
with a physical subdivision based at least in part on the RF
fingerprint and the environmental signature.
[0024] A storage device is provided. The storage device may include
machine-readable instructions that, when executed by a circuit,
cause the circuit to provide a semantic labeling controller, that
selectively generates at least one RF fingerprint using at least
one parameter from each of the plurality of received RF signals;
receives from the sensor at least one output signal that includes
data or information indicative of a physical subdivision external
to the portable electronic device; selectively generates at least
one environmental signature using at least one parameter from the
at least one sensor output signal; selectively groups data
representative of the at least one RF fingerprint and the data or
information representative of the at least one environmental
signature into a data cluster responsive to the RF fingerprint
meeting at least one threshold criteria; and autonomously proposes
a semantic label for logical association with a physical
subdivision based at least in part on the RF fingerprint and the
environmental signature.
[0025] FIG. 1 provides a schematic of an illustrative autonomous
location detection and semantic labeling system 100, in accordance
with at least one embodiment of the present disclosure. In
embodiments, a number of radio frequency (RF) transmitters
104A-104n (collectively, "RF transmitters 104") may each be located
in one of a respective number of locations 102A-102n (collectively
"locations 102"). Each RF transmitter 104A-104n generates a
respective RF signal 106A-106n (collectively "RF signal 106") that
includes a respective unique identifier 108A-108n and has a signal
strength 110A-110n (collectively "signal strength 110") that varies
inversely with the square of the distance from the originating RF
transmitter 104. A portable electronic device 122, such as a
smartphone or similar portable, processor-based, computing device
may be disposed at a location 124A at a location 112 that is
different from, but in the vicinity of, the locations 102. For
example, the portable electronic device 120 may be disposed in one
room 130A of a structure 112 that includes a number of subdivisions
130A-130n (collectively "subdivisions 130"), such as rooms,
offices, conference rooms, etc., that may be separated by full-or
partial-height partitions, walls, or other similar dividers. In
some implementations, an RF device 114 may be disposed in a
different room 130n within the structure 112.
[0026] The portable electronic device 122 may be located at a
physical location 124A, for example in subdivision 130A of
structure 112. At physical location 124A places the portable
electronic device a distance from each of RF devices 104A-104n. At
physical location 124A, the unique identifiers 108A-108n and the
signal strengths 110A-110n of RF signals 106A-106n and the unique
identifier 118 and the signal strength 120 of RF signal 116 form a
data set or data cluster that contains a number of values or an "RF
fingerprint" unique to physical location 124A. In a like manner,
ANY physical location 124A-n within the system 100 may be
identified by a unique RF fingerprint.
[0027] As the portable electronic device 122 is moved to different
physical locations (e.g., different subdivisions 130A-130n) in the
structure 112, the RF fingerprint (i.e., the data set containing
the unique identifier and signal strength of each RF signal
104A-104n and 114) at the portable electronic device will change.
Indeed, as the portable electronic device 120 moves about a single
subdivision 130 within a structure 112, the RF fingerprint detected
by the portable electronic device 122 will change. Such variances
in RF fingerprints make mapping the extent of a single subdivision
130 difficult, tedious, and time consuming.
[0028] In implementations, the portable electronic device 122 may
include a number of sensors. The sensors carried by the portable
electronic device 122 may collect information about all or a
portion of the external or ambient environment about the portable
electronic device 122. Such information may include, but is not
limited to, an ambient audio level in the environment, an ambient
light level in the environment, a movement of the portable
electronic device in the environment, and/or one or more images of
the environment. In implementations, the sensors carried by the
portable electronic device 122 may provide a respective output
signal that includes data or information representative of one or
more aspects of the ambient environment about the portable
electronic device 122.
[0029] In implementations, each subdivision 130 within structure
112 may include elements 140 that uniquely characterize and/or
identify the ambient environment within the respective subdivision
130. Each of these elements 140 may include one or more
environmental signatures 142A-142n (collectively "environmental
signature 142") that characterize the element 140 and which may
assist in identification of the subdivision 130. These
environmental signatures 142 may include unique colors, unique
shapes, unique sounds, or combinations thereof. For example, a
subdivision 130, such as a kitchen, may include various appliances
(e.g., refrigerator140A, air conditioner 140B, garbage disposal
140n) that produce a recognizable, reproducible, and identifiable
environmental signature 142A-142n. In such instances, an audio
sensor carried by the portable electronic device 122 may capture
data or information capable of characterizing a particular
subdivision 130 within the structure 112. For example, a microphone
carried by the portable electronic device 122 may output an audio
signal that includes data or information representative of the
characteristic audio of a refrigerator running or a garbage
disposal in operation in a subdivision 130 such as a kitchen. Such
audio information may be captured at some or all of the physical
locations 124 within the kitchen subdivision 130.
[0030] Combining received RF fingerprint data with sensor output
data generated by one or more sensors carried by the portable
electronic device 122 may form a data cluster. In embodiments, the
portable electronic device 122 may selectively generate one or more
data clusters by combining some or all of the received RF
fingerprint data with some or all of the data or information
included in received sensor output data. For each physical location
124 within a subdivision 130, the RF fingerprint may vary, however
sufficient similarity may exist between the data or information
included in the portable electronic device sensor output data to
permit the conclusion that all of the data clusters having
different RF fingerprints but similar sensor output data represent
different physical locations 124 within a common subdivision
130.
[0031] For example, a number of RF fingerprints will exist at
different physical locations 124 within a subdivision 130 such as a
kitchen, however, the characteristic sounds of a running
refrigerator or running garbage disposal may be heard throughout
the kitchen subdivision 130. In such an instance, it may be
concluded that while the RF fingerprints logically associated with
some or all of the data clusters differ (i.e., represent different
physical locations 124) and some or all of the data clusters share
sufficiently similar sensor output data (i.e., audio signal that
includes a refrigerator running or a garbage disposal operating)
represent physical locations 124 within a common subdivision 130
(i.e., the kitchen). In such an instance, the portable electronic
device 120 may propose a semantic label such as "kitchen" for
logical association with the "virtual" subdivision defined by each
of the data clusters. In other words, the portable electronic
device is able to assign a semantic label to a collection of data
clusters defined by one or more physical locations 124 in the
structure 112.
[0032] Although depicted as structures in FIG. 1, the locations 102
may include any number and/or combination of indoor locations and
outdoor locations. For example, a portable electronic device 120
may be positioned in a physical location where RF signals 106 from
multiple RF transmitters 104 are received. Such may include a
mixture of RF signals such as: an unsecure outdoor Wi-Fi connection
in a public park, an unsecure indoor Wi-Fi connection in a library,
and a secure indoor Wi-Fi connection in a private home. Each of the
locations 102 contains one or more RF transmitters 104 that
generate an RF signal 106 that contains at least a unique
identifier 108 that is logically associated or associable with the
originating RF transmitter 104.
[0033] The RF transmitter 104 may include any number and/or
combination of current and/or future transmitters, receivers, or
transceivers capable of producing, generating, and emitting an RF
signal 106 that includes a unique identifier 108. In some
implementations, the RF transmitter 104 may include one or more
wireless routers, such as one or more IEEE 802.11 (latest version)
compliant wireless routers. In some implementations, the RF
transmitter 104 may include one or more wired or wireless modems,
such as one or more wired cable modems, one or more wired digital
subscriber line (DSL) modems, or one or more satellite connected
modems. In some implementations, the RF transmitter 104 may include
one or more portable electronic devices capable of operation in a
"wireless hotspot" or "hotspot" mode.
[0034] Each RF transmitter 104 may produce, generate, and/or emit
an RF signal 106 that includes a unique identifier 108 and has a
signal strength 110 that varies somewhere between the inverse of
the square of the distance (1/d.sup.2) and the inverse of the
fifth-power of the distance (1/d.sup.5) from the respective RF
transmitter. In embodiments, each RF transmitter 104 may form a
portion of a network such as a packet switched network using
frames, packets, or similar discrete logical structures that
include a header or similar logical format that includes data
indicative of an identifier unique to the respective RF
transmitter. In such embodiments, the unique identifier 108 may
include data indicative of a network interface card (NIC)
identifier. In some implementations, the unique identifier may
include data indicative of a media access card (MAC) address. In
some implementations, the RF transmitter 104 may include a
transmitter or transceiver compliant with any current or future
IEEE 802.11 (Wi-Fi) standard. In some implementations, the RF
transmitter 104 may include a transmitter or transceiver compliant
with any current or future BLUETOOTH.RTM. standard. In some
implementations, the RF transmitter 104 may include a transmitter
or transceiver compliant with any current or future Near Field
Communication (NFC) standard. In some implementations, the RF
transmitter 104 may include any number or combination of
transmitters or transceivers compliant with any current or future
wireless digital data transmission standard.
[0035] The structure 112 may include any structure having one or
more subdivisions 130A-130n. In some implementations, the structure
112 may include an enclosed structure such as a house, commercial
building, shopping mall, or any other closed structure. In some
implementations, the structure 112 may include an open or semi-open
structure such as an outdoor park, amphitheater, or any other
partially or completely open structure in a defined area. A single
structure 112 may include or incorporate any number of physical
locations 124A-124n. A single structure 112 may include or
incorporate any number of subdivisions 130. A single subdivision
130 may include any number of physical locations 124.
[0036] For example, a structure 112 such as an office building may
include tens or even hundreds of subdivision 130 such as offices
and conference rooms. Each of the subdivisions 130 in the structure
112 may include a number of physical locations 124. In another
example, a structure 112 may include a public park that includes a
number of subdivisions 130 such as a children's playground, a
tennis court, a swimming pool, and a nature area. Each of the
subdivisions 130 in the structure 112 may include a number of
physical locations 124.
[0037] FIG. 2 depicts an illustrative subdivision 130 that includes
multiple elements 140 and multiple portable electronic devices
122A-122n disposed in different physical locations 124A-124n
throughout the subdivision 130, in accordance with at least one
embodiment of the present disclosure. The subdivision 130 may
include any indoor or outdoor location. As depicted in FIG. 2,
portable electronic devices 122A and 122B are disposed within the
subdivision 130 at physical locations 124A and 124B, respectively.
Portable electronic devices 122A and 122B communicate a data
cluster 212A, 212B and a logically associated sematic label 210A,
210B to portable electronic device 122n. Data clusters 212A and
212B include respective RF fingerprints 214A, 214B (collectively,
"RF fingerprints 214") and respective environmental fingerprints
216A, 216B. In one example, the environmental signature 216A and
216B (collectively, "environmental signatures 216") provided by the
portable electronic devices 122A and 122B to the portable
electronic device 122n may include a number of environmental
signatures 142A-142n originating from a respective number of
environmental elements 140A-140n disposed in, on, about, or around
the subdivision 130.
[0038] In one embodiment, portable electronic devices 122A and 122B
may be disposed in a kitchen and the environmental elements may
include items typically encountered in a kitchen, such as a
refrigerator 140A, a garbage disposal 140B, and a range hood 140C.
The environmental signatures 142A-142C may be generated by the
respective environmental element 140A-140C. In such an embodiment,
portable electronic devices 122A and 122B may communicate data
clusters 212A and 212B, respectively, to portable electronic device
122n. Portable electronic device 122n may receive data clusters
212A and 212B and compare the environmental signature 142A and 142B
with the locally collected environmental signature 142n. If the
locally collected environmental signature 142n compares favorably
with (i.e., within a defined limit or tolerance) environmental
signatures 142A and 142B, portable electronic device 122n may
conclude that physical location 124n lies within the same
subdivision 130 as physical locations 124A and 124B, and may
suggest or propose the use of the semantic label 210A and 210B
"KITCHEN" provided by portable electronic devices 122A and
122B.
[0039] Such communication between portable electronic device 122n
and portable electronic devices 122A and 122B beneficially improves
the accuracy and speed of determining an appropriate semantic label
using one or more virtual room data clusters collected by portable
electronic device 122n. This merging of location data or
information provided by the RF fingerprint with a locally obtained
environmental signature 216 provides significant improvements in
accuracy and semantic labeling when compared to systems and methods
reliant on RF fingerprinting alone.
[0040] FIG. 3 and the following discussion provide a brief, general
description of the components forming an illustrative system 300
that includes one or more semantic labeling controller(s) 312 in
which the various illustrated embodiments of the portable
electronic device 122 may be implemented. Some embodiments may
include machine-readable or computer-executable instruction sets,
such as program application modules, objects, or macros being
executed by the one or more semantic labeling controller(s) 312.
Some embodiments may include circuitry implemented in the form of
hard-wired circuitry, semiconductor circuitry, microprocessors,
controllers, or similar devices that provide radio frequency (RF)
fingerprint circuitry 314, environmental signature circuitry 316,
data grouping circuitry 318, semantic labeling circuitry 320,
device profile selection circuitry 322, and similar.
[0041] Those skilled in the relevant art will appreciate that the
illustrated embodiments as well as other embodiments can be
practiced with other circuit-based device configurations, including
portable electronic or handheld electronic devices, for instance
smartphones, portable computers, wearable computers,
microprocessor-based or programmable consumer electronics, personal
computers ("PCs"), network PCs, minicomputers, mainframe computers,
and the like. The embodiments can be practiced in distributed
computing environments where tasks or modules are performed by
remote processing devices, which are linked through a
communications network. In a distributed computing environment,
program modules may be located in both local and remote memory
storage devices.
[0042] The portable electronic device 122 may include any number of
circuits 310, each of which may include a variety of electronic
and/or semiconductor components that are disposed partially or
wholly in a wearable computer, portable computing device, personal
digital assistant, personal computer, or other similar current or
future processor-based devices and/or systems capable of executing
machine-readable instructions. The one or more circuits 310 may be
interconnected with, electrically coupled, and/or communicably
coupled to various components within the portable electronic device
122 via one or more communications links 306. As depicted in FIG.
3, the one or more circuits 310 may be apportioned or allocated to
one or more controllers 312, RF fingerprint generation circuitry
314, environmental signature circuitry 316, data grouping circuitry
318, semantic labeling circuitry 320, and device profile selection
circuitry 322.
[0043] As depicted in FIG. 3, system components such as a system
memory 322 may be communicably coupled to the one or more circuits
310 via the one or more communications links 306. The portable
electronic device 122 may, at times, be referred to in the singular
herein, but this is not intended to limit the embodiments to a
single system, since in certain embodiments, there will be more
than one portable electronic device 122 or other networked systems,
circuits, or devices involved.
[0044] The one or more circuits 310 may include any number, type,
or combination of conductors, insulators, electrical devices,
and/or semiconductor components. At times, the one or more circuits
310 may be implemented in whole or in part in the form of
semiconductor devices such as diodes, transistors, inductors,
capacitors, and resistors. Such an implementation may include, but
is not limited to any current or future developed single-or
multi-core processor or microprocessor, such as: one or more
systems on a chip (SOCs); one or more central processing units
(CPUs); one or more digital signal processors (DSPs); one or more
graphics processing units (GPUs); one or more application-specific
integrated circuits (ASICs), one or more field programmable gate
arrays (FPGAs), and the like. Unless described otherwise, the
construction and operation of the various blocks shown in FIG. 3
are of conventional design. As a result, such blocks need not be
described in further detail herein, as they will be understood by
those skilled in the relevant art. The one or more communications
links 306 that interconnects at least some of the components may
employ any known serial or parallel bus structures or
architectures.
[0045] The system memory 322 may include read-only memory ("ROM")
324 and random access memory ("RAM") 326. A portion of the ROM 324
may contain a basic input/output system ("BIOS") 328. The BIOS 328
may provide basic functionality to the portable electronic device
122. For example, by causing, in some implementations, the one or
more circuits 310 to load one or more machine-readable instruction
sets that cause the one or more circuits 310 to provide and
function as a particular and specialized semantic labeling
controller 312. In another example, the BIOS 328 may cause a
general purpose circuit 310, processor, microprocessor, or similar
device to execute machine-readable instruction sets that cause the
general purpose circuit 310 to transform into one or more
specialized or particular machines and/or circuitry such as the RF
fingerprint generation circuitry 314, environmental signature
circuitry 316, data grouping circuitry 318, semantic labeling
circuitry 320, and device profile selection circuitry 322.
[0046] The portable electronic device 122 may include one or more
communicably coupled, non-transitory, data storage devices 332. The
one or more data storage devices 332 may include any number and/or
combination of any current or future developed non-transitory
storage devices and/or memory. Non-limiting examples of such
non-transitory, data storage devices 332 may include, but are not
limited to one or more magnetic storage devices, one or more
optical storage devices, one or more solid-state electromagnetic
storage devices, one or more electroresistive storage devices, one
or more molecular storage devices, one or more quantum storage
devices, or various combinations thereof.
[0047] The one or more storage devices 332 may include interfaces
or controllers (not shown) communicatively coupling the respective
storage device(s) 332 to the one or more communications links 306,
as is known by those skilled in the art. The one or more storage
devices 332 may contain machine-readable instruction sets, data
structures, program modules, and other data useful to the portable
electronic device 122. In some instances, one or more external
storage devices 330 may be communicably coupled to the portable
electronic device 122. Such external storage devices 330 may be
local to the portable electronic device 122 or remote from the
portable electronic device 122. In one example, the one or more
external storage devices 330 may include one or more remote server
based storage devices or "cloud" storage devices that are
bi-directionally communicably coupled to the portable electronic
device 122 via one or more networks, such as one or more wireless
networks (e.g., IEEE 802.11 "Wi-Fi"), one or more wired networks
(e.g., IEEE 802.3"Ethernet"), or combinations thereof.
[0048] The portable electronic device 122 may include a number of
sensors 380 that are communicably coupled to the semantic labeling
controller 312 and/or the environmental signature circuit 316 via
the one or more communications links 306. In some implementations,
the sensors 380 may include some or all of: one or more audio input
sensors 382 (e.g., one or more microphones or similar); one or more
image sensors 384 (e.g., one or more charge coupled device (CCD) or
complementary metal oxide semiconductor (CMOS) image sensors,
ambient light level sensors, or similar); one or more infrared
illuminators/infrared image acquisition sensors 386; one or more
movement or motion based sensors 388, or combinations thereof. In
at least some implementations, such sensors 380 may autonomously
collect environmental data or information on a regular, irregular,
periodic, or aperiodic basis.
[0049] Machine-readable instruction sets and/or applications 338
and housekeeping instruction sets 340 may be stored or otherwise
retained in whole or in part in the system memory 324. Such
instruction sets may be transferred from one or more storage
devices 332 and/or one or more external storage devices 332 and
stored in the system memory 324 in whole or in part for execution
by the one or more circuits 310. The machine-readable instruction
sets 338 may include instructions and/or logic providing the
semantic labeling functions and capabilities described herein.
[0050] For example, one or more applications 338 may cause the RF
fingerprint generation circuitry 314 to collect data representative
of one or more parameters associated with each of a plurality of RF
signals received by the portable electronic device 122. The one or
more applications 338 may further cause the RF fingerprint
generation circuitry 314 to generate a signal that includes the
collected data representative of one or more parameters associated
with each of a plurality of RF signals received by the portable
electronic device 122.
[0051] In some implementations, the RF fingerprint generation
circuitry 314 may collect data from a plurality of RF signals
received within at least one defined frequency band. Such frequency
bands may include, but are not limited to, an IEEE 802.11 (WiFi)
frequency band, a BLUETOOTH.RTM. frequency band, a Near Field
Communication frequency band, or combinations thereof. In some
implementations, the RF fingerprint generation circuitry 314 may
generate a signal containing data associated with each of the
received RF signals--this data may include, but is not limited to,
one or more of the following: a signal strength, a signal
frequency, a unique signal identifier, a unique signal transmitter
identifier, a signal phase, or a signal direction. The one or more
applications 338 may further cause the RF fingerprint generation
circuitry 314 to combine the collected RF signal information or
data into an RF fingerprint 214 that may be logically associated
with a particular physical or geographic location 124.
[0052] The one or more applications 338 may additionally cause the
RF fingerprint generation circuitry 314 to collect RF samples at
defined intervals which may be fixed, periodic, or based on the
occurrence or lack of occurrence of one or more defined events. For
example, the one or more applications 338 may cause the RF
fingerprint generation circuitry 314 to collect RF samples at a
slower rate when the portable electronic device 122 remains
stationary or when the portable electronic device 122 remains
stationary for a time greater than a defined temporal interval.
[0053] The one or more applications 338 may further cause the RF
fingerprint generation circuitry 314 to discard collected RF
samples or to not collect RF samples based on an occurrence or a
lack of an occurrence of one or more defined events. For example,
the one or more machine-readable instruction sets 338 may cause RF
fingerprint generation circuitry to halt collection of RF samples
when the portable electronic device 122 is moving or when the
portable electronic device 122 is moving at a rate or velocity
greater than a defined value.
[0054] The one or more applications 338 may cause the environmental
signature circuitry 316 to collect environmental data or
information from the one or more sensors 380. The one or more
applications 338 may cause the environmental signature circuitry
316 to further generate an environmental signature signal that
includes data indicative of one or more parameters of the
environment about the portable electronic device 122.
[0055] Such environmental data may include audio information
collected using one or more audio sensors 382 (e.g., one or more
microphones); one or more image sensors 384 (e.g., one or more
image acquisition devices sensitive in one or more spectral bands
such as the visible spectral band of 380 nanometers to 760
nanometers, or the near infrared spectral band greater than 760
nm). Such environmental data may include movement, motion,
orientation, and/or acceleration information or data collected from
one or more MEMS devices, such as one or more accelerometers or
gyroscopic devices coupled to a portable electronic device. Such
environmental data may include geolocation data, such as global
positioning system data. In embodiments, such environmental data or
information may be collected contemporaneously with the collection
of RF signal data by the RF fingerprint generation circuitry. In
other embodiments, such environmental data or information may be
collected at different times than the collection of RF signal
data.
[0056] The one or more applications 338 may cause the data grouping
circuitry 316 to generate one or more data clusters 212 by
selectively combining one or more RF fingerprints 214 received from
the RF fingerprint generation circuitry 314 with one or more
environmental signatures 216 received from the environmental
signature circuits 316. In some implementations, the one or more
applications 338 may cause the portable electronic device 122 to
logically associate, on at least an interim basis, such data
clusters 212 with a "virtual" room or similar logical entity prior
to establishing a logical association with a defined semantic
label.
[0057] The one or more applications 338 may cause the semantic
labeling circuitry 320 to autonomously analyze some or all of the
data clusters and autonomously generate a proposed semantic label
210 for logical association with at least one of: the data cluster
212, the RF fingerprint 214, and/or the environmental signature
216. In some implementations the one or more applications 338 may
prompt a device user to accept the proposed semantic label 210. In
some implementations, the one or more applications 338 may cause
the semantic labeling circuit 320 to autonomously analyze the data
clusters by comparing some or all of the environmental signatures
216 included in the data cluster to a plurality of known or
identified environmental signatures 216 that have been previously
logically associated with a respective semantic label 210.
[0058] The one or more applications 338 may cause the semantic
labeling circuitry 320 to autonomously populate an environmental
signature database, data structure, or data store with information
and/or data indicative of environmental signatures and their
logically associated semantic label. In some instances, the one or
more applications 338 may cause the semantic labeling circuitry 320
to exchange such environmental signature/semantic label information
with one or more other local portable electronic devices 122. In
some instances, the one or more applications 338 may cause the
semantic labeling circuitry 320 to communicate such environmental
signature/semantic label information to one or more remote servers
or cloud-based storage devices accessible to multiple portable
electronic devices 122. In such a manner, the systems and methods
described herein may beneficially "learn" new environmental
signatures 216 and provide more accurate sematic label suggestions
based on the environmental data or information received by the
semantic labeling circuitry 320.
[0059] The one or more applications 338 may cause the RF
fingerprint generator circuitry 314 and/or the environmental
signature circuitry 316 to selectively filter or otherwise apply
noise reduction techniques to the received RF signals or
information and/or the received environmental data or information.
In some implementations, the one or more applications 338 may cause
the environmental signature circuitry 316 to extract one or more
environmental signature(s) 216 from a noisy signal provided by one
or more sensor(s) 380. In some implementations, the one or more
applications 338 may cause the environmental signature circuitry
316 to perform an i-vector linear discriminant analysis
(i-vector/LDA) to take all acoustic features (Mel-frequency
Cepstral Coefficients/Perceptual Linear Prediction or MFCC/PLP) in
a segment of environmental data, such as an audio signal collected
using an audio input device 382. In some other implementations, the
MFCC/PLP or mel-filterbank outputs may cause the environmental
signature circuitry 316 to extract bottleneck features of a deep
neural network or convolutional neural network trained to
discriminate different sounds sources, and the bottleneck features
may be computed for every data frame which are averaged to produce
one vector for the segment. In embodiments, to account for the
random nature of sounds within the environment in a subdivision
130, the environmental data may be clustered into a small number of
centroids. Depending on local environmental conditions, a single
semantic label (e.g., "Kitchen," "Bedroom") may contain more than
one centroid to accommodate the inherent variability of the
environment within the subdivision. In embodiments, the one or more
applications 338 may cause the environmental signature circuitry
316 to infer a semantic label by determining a distance to the
nearest centroid.
[0060] The one or more applications 338 may cause the device
profile selection circuitry 322 to retrieve one or more device
parameters from a data store, data set, data base, or similar data
storage structure. In some implementations, the one or more
applications 338 may cause the device profile selection circuitry
322 to select the one or more device profiles based at least in
part on an RF fingerprint 214, an environmental signature 216, a
semantic label 210, or combinations thereof.
[0061] The portable electronic device 122 may include one or more
communicably coupled physical input devices 350, such as one or
more text entry devices 352 (e.g., keyboard), one or more pointing
devices 354 (e.g., mouse, trackball, touchscreen), and/or one or
more audio input devices 356. Such physical input devices 350 may
be used, for example, to provide, enter, or otherwise supply
commands (e.g., acknowledgements, selections, confirmations, and
similar) as well as information (e.g., acknowledgements, corrected
semantic labels, and similar) to the semantic labeling controller
312. The portable electronic device 122 may include one or more
communicably coupled physical output devices 360, such as one or
more visual output devices 362 (e.g., a display device), one or
more tactile output devices 364 (e.g., haptic feedback or similar),
one or more audio output devices 366, or any combination
thereof.
[0062] For convenience, the network interface 370, the one or more
circuits 310, the system memory 324, the physical input devices 350
and the physical output devices 360 are illustrated as
communicatively coupled to each other via the one or more
communications links 306, thereby providing connectivity between
the above-described components. In alternative embodiments, the
above-described components may be communicatively coupled in a
different manner than illustrated in FIG. 3. For example, one or
more of the above-described components may be directly coupled to
other components, or may be coupled to each other, via one or more
intermediary components (not shown). In some embodiments, the one
or more communications links 306 may be omitted and the components
are coupled directly to each other using suitable wired or wireless
connections.
[0063] FIG. 4 provides a high-level, block flow diagram of an
illustrative semantic labeling method 400 that may be implemented
on a portable electronic device 122, in accordance with at least
one embodiment of the present disclosure. The method 400 commences
with arrival at an identified location at 402. In embodiments, the
arrival may be detected by receiving data from a previously
identified source, for example a home Wi-Fi server. In embodiments,
arrival at a location, coupled with one or more defined physical
activities 404 causes the RF fingerprint generation circuitry 314
to scan 406 for the presence of other RF signals. In such
embodiments, such RF signal scanning may be instigated based on the
portable electronic device remaining in a stationary location or
within a defined distance of a physical location 124 for a time
that exceeds a defined threshold.
[0064] In some implementations, the RF fingerprint generation
circuitry 314 may halt the collection of RF signal data or
information at 408, for example when the RF fingerprint generation
circuitry 314 determines the portable electronic device 112 is in
motion between physical locations 124. In some implementations, the
RF fingerprint generation circuitry 314 may filter or otherwise
condition the collected RF signals based at least in part on the
movement or motion of the portable electronic device 122 within a
subdivision 130 at 408.
[0065] In some implementations, the RF fingerprint generation
circuitry 314 may aggregate the filtered RF signals and extract
features characterizing the received RF signals to provide an RF
fingerprint 214 at 412. After collecting a sufficient number of RF
signals and extracting the features characterizing the received RF
signals, the RF fingerprint generation circuitry 314 may commence a
background task that includes unsupervised clustering of the
features into one or more data clusters 212. To reduce noise and
signal fluctuation in the received RF data the RF fingerprint
generation circuitry 314 may apply a principal component analysis
(PCA) or other dimensionality reduction and de-noising techniques
to the features characterizing the received RF signals prior to
clustering at 414.
[0066] In embodiments, the semantic labeling circuitry 322 may
opportunistically prompt for user annotation and feedback at 410.
In some implementations, the semantic labeling circuitry 322 may
opportunistically prompt for user annotation and feedback based at
least in part on the motion or movement of the portable electronic
device 122. In at least some implementations, the user supplied
input may be used to merge subdivisions and to map virtual
subdivisions to physical subdivisions at 416 by logically
associating meaningful semantic labels to the collected RF data. In
embodiments, the semantic labeling circuitry 320 may request such
opportunistic feedback when the user is detected in an idle state
(i.e., a non-moving state). Such opportunistic feedback may include
providing a correct label for the subdivision 130 currently
occupied by the user; acknowledging that the autonomously selected
semantic label 210 proposed by the semantic labeling circuitry 320
is correct; or, correcting an incorrect, autonomously selected,
semantic label 210 proposed by the semantic labeling circuitry
320.
[0067] In some implementations, the environmental signature
circuitry 316 may collect ambient environmental signatures at 420.
The semantic labeling circuitry 320 may logically associate one or
more environmental signatures 216, for example one or more
environmental audio signatures, with a physical subdivision 130,
such as a physical room, at 416. In some implementations, the
semantic labeling circuitry 320 may identify one or more virtual
subdivisions by matching or otherwise correlating one or more
environmental signatures 216 collected from the virtual subdivision
with one or more environmental signatures 216 logically associated
with a physical subdivision 130.
[0068] In some implementations, the semantic labeling circuitry 320
may logically associate an environmental signature 216 with a
physical subdivision 130. In order to provide timely and
appropriate requests for user annotation of virtual subdivisions or
confirmation of an autonomously generated semantic label 210
proposed by the semantic labeling controller 312 (e.g., the user
and the portable electric device 122 are stationary), the semantic
labeling circuitry 320 may monitor the physical activity of the
user at 424 or the movement of the portable electronic device 122.
Using a stationary period filter at 422 that determines when the
physical activity of the user or the movement of the of the
portable electronic device 122 meets one or more defined criteria
the semantic labeling circuitry 320 may autonomously propose a
semantic label for one or more virtual subdivisions based at least
in part on the environmental signatures 216 logically associated
with physical subdivisions 130.
[0069] In at least some implementations, the semantic labeling
circuitry 320 may be trained at 418 based on semantic labeling
information provided by the user of the portable electronic device
122 and/or user confirmation of a semantic label autonomously
proposed by the semantic labeling circuitry 320. The trained
semantic labeling circuitry 320 may, in some implementations, use a
classifier or similar logical structure, such as a decision tree
classifier at 432, to propose future semantic labeling based on
data or information such as the RF fingerprint 214 and/or the
environmental signature 216 logically associated with a virtual
subdivision. In embodiments, such training may also provide
feedback on the performance of the semantic labeling circuitry 320.
Such feedback may beneficially be used to identify
"high-performing" semantic labeling circuitry 320 able to reliably
and accurately identify appropriate semantic labels 210 for virtual
subdivisions. Such high performing semantic labeling circuitry 320
may then reliably and accurately autonomously logically associate
semantic labels 210 with virtual subdivisions in the absence of
user input. For example, if a certain population of RF fingerprints
in a virtual subdivision may be classified with high confidence as
a known physical subdivision, then the virtual subdivision may
share the room label and be correctly mapped into the corresponding
physical subdivision.
[0070] FIG. 5 provides a high-level block flow diagram of an
example environmental signature content extraction method 500 that
may be used with the semantic labeling method 400 in FIG. 4, in
accordance with at least one embodiment of the present disclosure.
Although FIG. 5 is discussed in terms of an acoustic environmental
signature, those of ordinary skill in the art will appreciate the
applicability of the systems and methods disclosed herein to other
forms and types of environmental signatures.216 that do not rely
upon an acoustic environmental signature.
[0071] In embodiments, the environmental signature 216 or acoustic
signature of a subdivision 130 such as a room may primarily be
characterized by the low energy or low frequency components rather
than the high energy or high frequency components. Thus, low
frequency sounds such as HVAC system operation, outside road
traffic, and refrigerator hum are environmental signatures 216 that
may be characterized by the environmental signature circuitry
316.
[0072] At 502, the environmental signature circuitry 316 collects
data representative of the ambient audio environment about the
portable electronic device 122. The collected audio data may
include both high frequency components (e.g., voices, television
audio, and similar) as well as low frequency components (e.g., HVAC
hum, refrigerator running, exterior traffic, and similar). In some
implementations, the environmental signature circuitry 316 may
selectively extract all or a portion of the low frequency
components present in the ambient audio data. The environmental
signature circuitry 316 may collect such low frequency components
of the environmental signature 216 for a defined time period, such
as 15 seconds, 30 seconds, 45 seconds, or 60 seconds.
[0073] At 504, the environmental signature circuitry 316 may
extract features from the collected environmental signature 216. In
embodiments, during the feature extraction at 504, the
environmental signature circuitry 316 may target desired properties
such as providing a compact representation of the environmental
signature 216 and discriminating between frequencies included in
the environmental signature 216 by positioning the frequencies far
apart in a feature space. In some implementations, the desired
properties of the environmental signature 216 may be targeted using
an i-vector/DLA and/or the bottleneck features of a deep neural
network (DNN). The i-vector collects all environmental signatures
in the collected environmental data or information and produces a
single vector; the bottleneck feature may then be computed for
every set of environmental signature data or information to produce
a segment of the single vector.
[0074] At 506, the data grouping circuitry 318 clusters the
environmental signatures 216 associated with a single virtual
subdivision into a number of centroids. One advantage of the
systems and methods described herein is the opportunity to prompt
for user feedback to assist the semantic labeling controller 312 in
merging and purging subdivisions.
[0075] The semantic labeling controller 312 disclosed herein
beneficially fuses RF fingerprint data 214 with environmental
signature data 216 to capture the context of a subdivision 130 and
leverages the fusion of the RF and environmental data to
autonomously propagate semantic labels or merge subdivisions 130.
For example, the environmental signature 216 of a subdivision 130
is likely unique across different subdivisions 130 and may be used
by the semantic labeling controller 312 as a signature to identify
different subdivisions 130. Virtual subdivisions discovered by the
semantic labeling controller 312 that share a similar environmental
signature 216 with a previously identified physical subdivision 130
may then merged (e.g., based on proximity data gathered from the RF
fingerprint 214) or labeled autonomously by the semantic labeling
controller 312 without requiring user input. Similarly, by fusing
environmental signature data such as accelerometer data, different
virtual subdivisions detected while the portable electronic device
122 is maintained stationary may be autonomously merged or labeled
accordingly if the time the portable electronic device 122 remains
stationary overlaps with the detection of a physical subdivision
130.
[0076] FIG. 6 presents a high-level logic flow diagram of an
illustrative semantic labeling method 600 that may be implemented
using one or more semantic labeling controllers 312 such as those
described in detail with regard to FIGS. 1 through 5, in accordance
with at least one embodiment of the present disclosure. The method
600 commences at 602.
[0077] At 604, the RF fingerprint generation circuitry 314
selectively detects a number of radio frequency (RF) signals 106.
Given the receipt of a plurality of RF signals 106 by the RF
fingerprint generation circuitry 314, a physical location 124 may
be identified based on a respective unique identifier 108 carried
by and logically associated with each of the RF signals 106 and a
respective signal strength of each of the RF signals 106. The RF
fingerprint generation circuitry 314 may filter, process, or
otherwise condition some or all of the received RF signals 106 to
provide an RF fingerprint 214 that is unique to a particular
physical location 124 within a three dimensional space.
[0078] In some implementations, the RF fingerprint generation
circuitry 314 may selectively sample the received RF signals on an
intermittent, continuous, periodic, or aperiodic basis. In some
implementations, such sampling may be based on one or more
environmental parameters, for example whether the portable
electronic device 122 is in motion and/or the velocity of the
portable electronic device 122. In some implementations, the RF
fingerprint generation circuitry 314 autonomously determines an RF
signal sampling rate.
[0079] At 606, the environmental signature circuitry 316 receives
one or more output signals provided by one or more environmental
sensors carried by the portable electronic device 122. In
embodiments, the output signals provide an environmental signature
216 that characterizes the physical location 124 of the portable
electronic device 122. In some implementations, the environmental
signature circuitry 316 collects environmental signature data 216
at least partially contemporaneous with the receipt of RF
fingerprint data 214. In some implementations, the environmental
signature circuitry 316 collects environmental signature data 216
asynchronously with the receipt of RF fingerprint data 214.
[0080] In embodiments, the environmental sensors 380 may include
any number and/or combination of sensors capable of providing the
semantic labeling circuitry 320 with an output signal that includes
data or information related to either the portable electronic
device 122 or the environment surrounding the portable electronic
device 122. Such data or information may include, but is not
limited to, data or information related to: environmental audio,
environmental imaging, environmental light levels, environmental
humidity, environmental temperature, environmental pressure,
movement/motion/velocity of the portable electronic device 122,
acceleration of the portable electronic device 122,
physical/gyroscopic orientation of the portable electronic device
122, or combinations thereof.
[0081] At 608, the data grouping circuitry 318 selectively clusters
at least a portion of the received RF fingerprint data 214 with at
least a portion of the environmental signature data 216 to provide
a data cluster 212. In embodiments, the data grouping circuitry 318
may logically associate one or more data clusters 212 with the
physical location 124 of the portable electronic device 122 at the
time the RF fingerprint 214 and the environmental signature 216
were collected by the semantic labeling controller 312. Each data
cluster 212 uniquely identifies a particular physical location 124
based on the RF fingerprint 214 and the environmental signature 216
received by the semantic labeling circuitry 320.
[0082] At 610, the semantic labeling circuitry 320 may autonomously
propose a semantic label 210 for logical association with a
physical subdivision 130 based at least in part on the data cluster
212 generated by the semantic labeling controller 312 at a physical
location 124 within the respective physical subdivision 130. In
some implementations, the semantic labeling circuitry 320 may
autonomously select a proposed semantic label 210 for logical
association with a physical subdivision 130 based at least in part
on a favorable comparison between an environmental signature 216
collected by the environmental signature circuitry 316 and an entry
in a database or data store that logically relates that include
environmental signatures 216 with physical subdivisions 130. In
some implementations, the semantic labeling circuitry 320 may
autonomously select a proposed semantic label 210 for logical
association with a physical subdivision 130 based at least in part
on a favorable comparison between an RF fingerprint 214 collected
by the RF fingerprint generator circuitry 314 and an entry in a
database or data store that logically relates that include RF
fingerprints 214 with physical subdivisions 130. In some
implementations, the semantic labeling circuitry 320 may
autonomously select a proposed semantic label 210 for logical
association with a physical subdivision 130 based at least in part
on a favorable comparison between an RF fingerprint 214 and an
environmental signature 216 and an entry in a database or data
store that logically relates that includes RF fingerprints 214 and
environmental signatures 216 with physical subdivisions 130.
[0083] In some implementations, the database, data store, or
similar data structure that logically relates RF fingerprints 214
and environmental signatures 216 with physical subdivisions 130 may
be stored or otherwise retained in a storage device 332, 334
communicably coupled to the semantic labeling controller 312. In
some implementations, the storage device 332 may be disposed local
to the semantic labeling circuitry 320. In other implementations,
the storage device 332 may be disposed remote from the semantic
labeling circuitry 320 and may be communicably coupled via one or
more wireless or wired networks. The method 600 concludes at
612.
[0084] FIG. 7 presents a high-level logic flow diagram of an
illustrative semantic labeling method 700 in which the device
profile selection circuitry 322 autonomously selects a device
profile based at least in part on the semantic label 210 logically
associated with the subdivision 130, in accordance with at least
one embodiment of the present disclosure. The method 700 commences
at 702.
[0085] At 704, the device profile selection circuitry 322 directly
or indirectly selects a portable electronic device profile based at
least in part on the physical subdivision 130 logically associated
with the physical location 124 of the portable electronic device
122. For example, the device profile selection circuitry 322 may
select a device profile that reduces display brightness and reduces
audio output responsive to the portable electronic device 122 being
positioned in a physical location 124 that is logically associated
with a semantic label 210 "BEDROOM." The method 700 concludes
706.
[0086] FIG. 8 presents a high-level logic flow diagram of an
illustrative semantic labeling method 800 in which a portable
electronic device 122A receives data clusters 212B-212n from one or
more other portable electronic devices 122B-122n disposed in the
same physical subdivision 130, in accordance with at least one
embodiment of the present disclosure. The method 800 commences at
802.
[0087] At 804, the portable electronic device 122A detects the
presence of at least one other portable electronic device 122B-122n
located in the same physical subdivision 130.
[0088] At 806, the portable electronic device 122A wirelessly
exchanges data with at least some of the other portable electronic
devices 122B-122n located in the same physical subdivision 130. In
some implementations, such an exchange of information may occur in
the background and may be transparent to the user of the portable
electronic device 122A. In some implementations, such an exchange
of information may occur only after receiving authorization to
transmit or receive information is provided by the user of the
portable electronic device 122A.
[0089] In embodiments, the portable electronic device 122A may have
collected one or more RF fingerprints 214A and one or more
environmental signatures 216A, but the semantic labeling controller
312A may be unable to provide a proposed semantic label 210A for
logical association with the physical subdivision 130. The other
portable electronic devices 122B-122n may each communicate a
respective data cluster 212B-212n to the portable electronic device
122A. Each of the data clusters 212B-212n may include a respective
semantic label 210B-210n and a respective environmental signature
216B-216n.
[0090] At 808, the semantic labeling controller 312A in the
portable electronic device 122A may compare information, such as
the environmental signature 216B-216n included in each of the
received data clusters 212B-212n, with the environmental signature
216A received using the environmental sensors 380A on the portable
electronic device 122A.
[0091] At 810, responsive to a favorable comparison between the
environmental signature 216A received using the environmental
sensors 380A on the portable electronic device 122A and an
environmental signature 216B-216n included in a received data
cluster 212B-212n, the semantic labeling controller 312A may
logically associate the semantic label 210B in the respective data
cluster 212B with the physical subdivision 130. The method
concludes at 812.
[0092] FIG. 9 presents a high-level logic flow diagram of an
illustrative semantic labeling method 900 in which the semantic
labeling controller 312 autonomously adjusts the RF signal and/or
environmental signal sampling rate based on one or more physical
aspects of the portable electronic device 122, in accordance with
at least one embodiment of the present disclosure. The method 900
commences at 902.
[0093] At 904, the semantic labeling controller 312 autonomously
adjusts the RF signal sampling rate based at least in part on a
physical aspect of the portable electronic device 122. In
embodiments, the semantic labeling controller 312 may limit or even
halt the collection of RF signal data and/or environmental signal
data when the portable electronic device 122 is in motion. Such
motion may be detected, for example, using one or more gyroscopic
sensors or one or more accelerometers disposed in, on, or about the
portable electronic device 122. In embodiments, the semantic
labeling controller 312 may limit or even halt the collection of RF
signal data and/or environmental signal data when the portable
electronic device 122 is placed in a defined pose or position, such
as face-down on a surface. In embodiments, the semantic labeling
controller 312 may limit or even halt the collection of RF signal
data and/or environmental signal data when the portable electronic
device 122 is performing one or more defined functions, such as
when providing global positioning system (GPS) route guidance, or
when the portable electronic device 122 is coupled to a charging
source. Such autonomous, active, management of data collection by
the semantic labeling controller 312 may beneficially limit the
volume of data collected and stored in the portable electronic
device 122, thereby conserving memory in the portable electronic
device 122 as well as reducing noise within the signal and
variations in collected data. The method 900 concludes at 906.
[0094] "Circuitry", as used in any embodiment herein, may include,
for example, singly or in any combination, hardwired circuitry,
programmable circuitry, state machine circuitry, and/or firmware
that stores instructions executed by programmable circuitry. An app
may be embodied as code or instructions which may be executed on
programmable circuitry such as a host processor or other
programmable circuitry. A module, as used in any embodiment herein,
may be embodied as circuitry. The circuitry may be embodied as an
integrated circuit, such as an integrated circuit chip.
[0095] The following examples pertain to further embodiments. The
following examples of the present disclosure may comprise subject
material such as a device, a method, at least one machine-readable
medium for storing instructions that when executed cause a machine
to perform acts based on the method, means for performing acts
based on the method and/or a system for collecting radio frequency
(RF) fingerprints214 and environmental signatures 216 in a physical
location 124, analyzing the collected data and generating one or
more proposed semantic labels for logical association with the
physical location and with the RF fingerprint 214 and/or
environmental signature 216.
[0096] According to example 1, there is provided a semantic
labeling controller. The controller may include a transceiver to
receive a plurality of radio frequency (RF) signals within a
defined portion of the RF spectrum; a sensor to generate a signal
that includes data or information indicative of an ambient
environment in a physical subdivision external to the controller;
RF fingerprint generator circuitry coupled to the transceiver, the
RF fingerprint generator circuitry to receive a plurality of RF
signals and selectively generate an RF fingerprint that includes
data indicative of at least one parameter associated with each of
the received RF signals; environmental signature generator
circuitry coupled to the sensor, the environmental signature
generator circuitry to receive the signal that includes data or
information indicative of the ambient environment in the physical
subdivision and selectively generate an environmental signature
that includes data indicative of the ambient environment external
to the controller; data grouping circuitry coupled to the RF
fingerprint generator circuitry and to the environmental signal
generator circuitry to selectively group the RF fingerprint and the
environmental signature into a data cluster responsive to the RF
fingerprint meeting at least one threshold condition; and semantic
labeling circuitry coupled to the data grouping circuitry to
autonomously generate a proposed semantic label for logical
association with the physical subdivision based at least in part on
the RF fingerprint and the environmental signature included in the
data cluster.
[0097] Example 2 may include elements of example 1 where the
transceiver, in operation, may receive RF signals within at least a
portion of an RF spectrum defined by Bluetooth.RTM. or Near Field
Communication (NFC) standards.
[0098] Example 3 may include elements of example 1 where the
transceiver, in operation, receives RF signals within at least a
portion of an RF spectrum defined by IEEE 802.11 (WiFi)
standards.
[0099] Example 4 may include elements of example 3 where the RF
fingerprint generator circuitry selectively generates the RF
fingerprint using at least a unique identifier and a signal
strength logically associated with each respective one of the
plurality of received RF signals.
[0100] Example 5 may include elements of example 4 where the
semantic labeling circuitry further: generates a human perceptible
output that includes data indicative of the proposed semantic
label; receives input indicative of a confirmation of the proposed
semantic label; and responsive to receipt of the input indicative
of the confirmation of the proposed semantic label, logically
associates the confirmed semantic label with the physical
subdivision.
[0101] Example 6 may include elements of example 5 where the sensor
comprises at least one of: an audio input sensor, an ambient light
sensor, an accelerometer, or an image sensor.
[0102] Example 7 may include elements of example 6, and may
additionally include device profile selection circuitry coupled to
the semantic labeling circuitry to select a device profile based at
least in part on the identified physical subdivision.
[0103] Example 8 may include elements of example 6, and may
additionally include device detection circuitry coupled to the
semantic labeling circuitry and to the data grouping circuitry, the
device detection circuitry to: detect a presence of at least one
other local portable electronic device; receive a signal from the
at least one other local portable electronic device, the signal
including data representative of a semantic label and an identified
data cluster logically associated with the semantic label; compare
the received data representative of the identified data cluster
with the data cluster generated by the data grouping circuitry; and
responsive to a favorable comparison, logically associate the
received semantic label with the physical subdivision.
[0104] Example 9 may include elements of example 1 where the data
grouping circuitry selectively determines an RF signal sampling
rate based at least in part on at least one of data or information
included in the output signal from the sensor.
[0105] Example 10 may include elements of example 9 where the data
grouping circuitry selectively reduces the RF signal sampling rate
responsive to data or information included in the output signal of
the sensor indicative of the sensor being stationary.
[0106] Example 11 may include elements of example 9 where the data
grouping circuitry further selectively halts RF signal sampling
responsive to data or information included in the output signal of
the sensor indicative of the sensor being in motion.
[0107] Example 12 may include elements of example 1 where the data
grouping circuitry selectively groups the RF fingerprint and the
environmental signature into a data cluster responsive to the RF
fingerprint remaining within an RF fingerprint threshold for
greater than a defined temporal threshold period.
[0108] According to example 13, there is provided a semantic
labeling method. The method may include selectively generating, by
radio frequency (RF) fingerprint generator circuitry, an RF
fingerprint that includes data indicative of at least one parameter
associated with each of the received RF signals; generating, by a
sensor, a signal that includes data or information indicative of an
ambient environment in a physical subdivision external to the
controller; generating, by environmental signature generating
circuitry coupled to the sensor, an environmental signature that
includes data indicative of the ambient environment external to the
sensor; selectively grouping, by data grouping circuitry coupled to
the RF fingerprint generator circuitry and to the environmental
signal generator circuitry, the RF fingerprint and the
environmental signature into a data cluster responsive to the RF
fingerprint meeting at least one threshold condition; and
autonomously generating, by semantic labeling circuitry coupled to
the data grouping circuitry, a proposed semantic label for logical
association with the physical subdivision based at least in part on
the RF fingerprint and the environmental signature included in the
data cluster.
[0109] Example 14 may include elements of example 13 where
selectively generating an RF fingerprint that includes data
indicative of at least one parameter associated with each of the
received RF signals may include selectively generating, by the RF
fingerprint generator circuitry, an RF fingerprint using at least
one parameter from each of the plurality of received RF signals,
each of the plurality of received RF signals within at least a
portion of an RF spectrum that includes at least one of: a
Bluetooth.RTM. RF spectrum or Near Field Communication (NFC) RF
spectrum.
[0110] Example 15 may include elements of example 13 where
selectively generating an RF fingerprint that includes data
indicative of at least one parameter associated with each of the
received RF signals may include selectively generating, by the RF
fingerprint generator circuitry, an RF fingerprint using at least
one parameter from each of the plurality of received RF signals,
each of the plurality of received RF signals within at least a
portion of an RF spectrum that includes an IEEE 802.11 (WiFi) RF
spectrum.
[0111] Example 16 may include elements of example 15 where
selectively generating at least one RF fingerprint using at least
one parameter from each of the plurality of received RF signals may
include selectively generating, by the RF fingerprint generator
circuitry, at least one RF fingerprint using at least a unique
identifier and a signal strength logically associated with each
respective one of the plurality of received RF signals.
[0112] Example 17 may include elements of example 13 where
autonomously generating a proposed semantic label for logical
association with the physical subdivision based at least in part on
the RF fingerprint and the environmental signature included in the
data cluster may include generating, by the semantic labeling
circuitry, a human perceptible output that includes data indicative
of the proposed semantic label; receiving, by the semantic labeling
circuitry, an input indicative of a confirmation of the proposed
semantic label; and logically associating, by the semantic labeling
circuitry, the confirmed semantic label with the physical
subdivision responsive to receiving the input indicative of the
confirmation of the semantic label.
[0113] Example 18 may include elements of example 13 where
generating an environmental signature that includes data indicative
of the ambient environment external to the sensor may include
receiving, by a sensor, an environmental signature that includes
data or information indicative of at least one of: an audible event
external to the portable electronic device, an ambient light event
external to the portable electronic device, an ambient motion event
external to the portable electronic device, or an image of the
environment external to the portable electronic device.
[0114] Example 19 may include elements of example 13, and may
further include selecting, by device profile selection circuitry
coupled to the semantic labeling circuitry, a portable electronic
device profile based at least in part on the identified physical
subdivision.
[0115] Example 20 may include elements of example 13, and may
additionally include detecting, by device detection circuitry
coupled to the semantic labeling circuitry and to the data grouping
circuitry, a presence of at least one other local portable
electronic device; receiving, by the device detection circuitry, a
signal from the at least one other local portable electronic
device, the signal including data representative of a semantic
label and an identified data cluster logically associated with the
semantic label; comparing, by the device detection circuitry, the
received data representative of the identified data cluster with
the data cluster generated by the data grouping circuitry; and
responsive to a favorable comparison, logically associating, by the
device detection circuitry, the received semantic label with the
physical subdivision.
[0116] Example 21 may include elements of example 13 where
selectively generating at least one RF fingerprint using at least
one parameter from each of the plurality of received RF signals may
include selectively determining, by the data grouping circuitry, an
RF signal sampling rate based at least in part on at least one of
data or information included in an output signal from the
sensor.
[0117] Example 22 may include elements of example 21 where
selectively determining an RF signal sampling rate based at least
in part on at least one of data or information included in an
output signal from the sensor may include selectively reducing the
RF signal sampling rate, by the data grouping circuitry, responsive
to data or information included in the output signal of the sensor
indicative of the sensor being stationary.
[0118] Example 23 may include elements of example 21 where
selectively determining an RF signal sampling rate based at least
in part on at least one of data or information included in an
output signal from the sensor may include selectively halting RF
signal sampling, by the data grouping circuitry, responsive to data
or information included in the output signal of the sensor
indicative of the sensor being in motion.
[0119] According to example 24, there is provided an autonomous
semantic labeling system, may include a means for selectively
generating an RF fingerprint that includes data indicative of at
least one parameter associated with each of the received RF
signals; a means for generating a signal that includes data or
information indicative of an ambient environment in a physical
subdivision external to the system; a means for generating an
environmental signature that includes data indicative of the
ambient environment external to the sensor; a means for selectively
grouping the RF fingerprint and the environmental signature into a
data cluster responsive to the RF fingerprint meeting at least one
threshold condition; and a means for autonomously generating a
proposed semantic label for logical association with the physical
subdivision based at least in part on the RF fingerprint and the
environmental signature included in the data cluster.
[0120] Example 25 may include elements of example 24 where the
means for selectively generating an RF fingerprint that includes
data indicative of at least one parameter associated with each of
the received RF signals may include a means for selectively
generating the RF fingerprint using at least one parameter from
each of the plurality of received RF signals, each of the plurality
of received RF signals within at least a portion of an RF spectrum
that includes at least one of: a Bluetooth.RTM. RF spectrum or Near
Field Communication (NFC) RF spectrum.
[0121] Example 26 may include elements of example 24 where the
means for selectively generating an RF fingerprint that includes
data indicative of at least one parameter associated with each of
the received RF signals may include a means for selectively
generating the RF fingerprint using at least one parameter from
each of the plurality of received RF signals, each of the plurality
of received RF signals within at least a portion of an RF spectrum
that includes an IEEE 802.11 (WiFi) RF spectrum.
[0122] Example 27 may include elements of example 26 where the
means for selectively generating the RF fingerprint using at least
one parameter from each of the plurality of received RF signals may
include a means for selectively generating the RF fingerprint using
at least a unique identifier and a signal strength logically
associated with each respective one of the received RF signals.
[0123] Example 28 may include elements of example 27 where the
means for autonomously generating a proposed semantic label for
logical association with the physical subdivision based at least in
part on the RF fingerprint and the environmental signature included
in the data cluster may include a means for generating a human
perceptible output that includes data indicative of the proposed
semantic label; a means for receiving an input indicative of a
confirmation of the proposed semantic label; and a means for
logically associating the confirmed semantic label with the
physical subdivision responsive to receiving the input indicative
of the confirmation of the semantic label.
[0124] Example 29 may include elements of example 28 where the
means for generating a signal that includes data or information
indicative of an ambient environment in a physical subdivision
external to the system may include a means for generating a signal
that includes data or information indicative of an ambient
environment that includes at least one of: an ambient audible
event, an ambient light event, an ambient motion event, or an image
of the ambient environment.
[0125] Example 30 may include elements of example 29, and may
additionally include a means for selecting a device profile based
at least in part on the identified physical subdivision.
[0126] Example 31 may include elements of example 29, and may
additionally include a means for detecting a presence of at least
one other local portable electronic device; a means for receiving a
signal from the at least one other local portable electronic
device, the signal including data representative of a semantic
label and an identified data cluster logically associated with the
semantic label; a means for comparing the received data
representative of the identified data cluster with the data cluster
generated by the data grouping circuitry; and a means for,
responsive to a favorable comparison, logically associating the
received semantic label with the physical subdivision.
[0127] Example 32 may include elements of example 24 where the
means for selectively generating an RF fingerprint that includes
data indicative of at least one parameter associated with each of
the received RF signals may include a means for selectively
determining an RF signal sampling rate based at least in part on at
least one of data or information included in an output signal from
the sensor.
[0128] Example 33 may include elements of example 24 where the
means for selectively generating an RF fingerprint that includes
data indicative of at least one parameter associated with each of
the received RF signals may include a means for selectively
reducing the RF signal sampling rate responsive to data or
information included in the output signal of the sensor indicative
of the sensor being stationary.
[0129] Example 34 may include elements of example 24 where the
means for selectively determining an RF signal sampling rate based
at least in part on at least one of data or information included in
an output signal from the sensor may include a means for
selectively halting RF signal sampling responsive to data or
information included in the output signal of the sensor indicative
of the sensor being in motion.
[0130] According to example 35, there is provided a storage device
that includes machine-readable instructions that, when executed by
a circuit, cause the circuit to provide a semantic labeling
controller, that selectively generates an RF fingerprint that
includes data indicative of at least one parameter associated with
each of the received RF signals; generates a signal that includes
data or information indicative of an ambient environment in a
physical subdivision external to the controller; generates an
environmental signature that includes data indicative of the
ambient environment external to the sensor; selectively groups, the
RF fingerprint and the environmental signature into a data cluster
responsive to the RF fingerprint meeting at least one threshold
condition; and autonomously generates a proposed semantic label for
logical association with the physical subdivision based at least in
part on the RF fingerprint and the environmental signature included
in the data cluster.
[0131] Example 36 may include elements of example 35 where the
machine-readable instructions that cause the semantic labeling
controller to selectively generate an RF fingerprint that includes
data indicative of at least one parameter associated with each of
the received RF signals further cause the semantic labeling
controller to: selectively generate an RF fingerprint using at
least one parameter from each of the plurality of received RF
signals, each of the plurality of received RF signals within at
least a portion of an RF spectrum that includes at least one of: a
Bluetooth.RTM. RF spectrum or Near Field Communication (NFC) RF
spectrum.
[0132] Example 37 may include elements of example 35 where the
machine-readable instructions that cause the semantic labeling
controller to selectively generate an RF fingerprint that includes
data indicative of at least one parameter associated with each of
the received RF signals further cause the semantic labeling
controller to: selectively generate an RF fingerprint using at
least one parameter from each of the plurality of received RF
signals, each of the plurality of received RF signals within at
least a portion of an RF spectrum that includes an IEEE 802.11
(WiFi) RF spectrum.
[0133] Example 38 may include elements of example 37 where the
machine-readable instructions that cause the semantic labeling
controller to selectively generate at least one RF fingerprint
using at least one parameter from each of the plurality of received
RF signals further cause the semantic labeling controller to:
selectively generate at least one RF fingerprint using at least a
unique identifier and a signal strength logically associated with
each respective one of the plurality of received RF signals.
[0134] Example 39 may include elements of example 35 where the
machine-readable instructions that cause the semantic labeling
controller to autonomously generate a proposed semantic label for
logical association with the physical subdivision based at least in
part on the RF fingerprint and the environmental signature included
in the data cluster further cause the semantic labeling controller
to: generate a human perceptible output that includes data
indicative of the proposed semantic label; receive an input
indicative of a confirmation of the proposed semantic label; and
logically associate the confirmed semantic label with the physical
subdivision responsive to receiving the input indicative of the
confirmation of the semantic label.
[0135] Example 40 may include elements of example 35 where the
machine-readable instructions that cause the semantic labeling
controller to generating an environmental signature that includes
data indicative of the ambient environment external to the sensor
further cause the semantic labeling controller to receive an
environmental signature that includes data or information
indicative of at least one of: an audible event external to the
portable electronic device, an ambient light event external to the
portable electronic device, an ambient motion event external to the
portable electronic device, or an image of the environment external
to the portable electronic device.
[0136] Example 41 may include elements of example 35, and may
additionally include machine-readable instructions that cause the
semantic labeling controller to select a portable electronic device
profile based at least in part on the identified physical
subdivision.
[0137] Example 42 may include elements of example 35, and may
additionally include machine-readable instructions that cause the
semantic labeling controller to: detecting a presence of at least
one other local portable electronic device; receiving a signal from
the at least one other local portable electronic device, the signal
including data representative of a semantic label and an identified
data cluster logically associated with the semantic label;
comparing the received data representative of the identified data
cluster with the data cluster generated by the data grouping
circuitry; and responsive to a favorable comparison, logically
associating the received semantic label with the physical
subdivision.
[0138] Example 43 may include elements of example 35 where the
machine-readable instructions that cause the semantic labeling
controller to selectively generating at least one RF fingerprint
using at least one parameter from each of the plurality of received
RF signals sensor further cause the semantic labeling controller to
selectively determine an RF signal sampling rate based at least in
part on at least one of data or information included in an output
signal from the sensor.
[0139] Example 44 may include elements of example 43 where the
machine-readable instructions that cause the semantic labeling
controller to selectively determining an RF signal sampling rate
based at least in part on at least one of data or information
included in an output signal from the sensor further cause the
semantic labeling controller to selectively reduce the RF signal
sampling rate, by the data grouping circuitry, responsive to data
or information included in the output signal of the sensor
indicative of the sensor being stationary.
[0140] Example 45 may include elements of example 43 where the
machine-readable instructions that cause the semantic labeling
controller to selectively determining an RF signal sampling rate
based at least in part on at least one of data or information
included in an output signal from the sensor further cause the
semantic labeling controller to selectively halt RF signal
sampling, by the data grouping circuitry, responsive to data or
information included in the output signal of the sensor indicative
of the sensor being in motion.
[0141] According to example 46, there is provided a system for
autonomous semantic labeling of spaces, the system being arranged
to perform the method of any of claims 13 through 23.
[0142] According to example 47, there is provided a chipset
arranged to perform the method of any of claims 13 through 23.
[0143] According to example 48, there is provided at least one
machine readable medium comprising a plurality of instructions
that, in response to be being executed on a computing device, cause
the computing device to carry out the method according to any of
claims 13 through 23.
[0144] According to example 49, there is provided a device
configured for autonomous labeling of spaces with semantic labels,
the device being arranged to perform the method of any of claims 13
through 23.
[0145] As used in any embodiment herein, the term "module" may
refer to software, firmware and/or circuitry configured to perform
any of the aforementioned operations. Software may be embodied as a
software package, code, instructions, instruction sets and/or data
recorded on non-transitory computer readable storage mediums.
Firmware may be embodied as code, instructions or instruction sets
and/or data that are hard-coded (e.g., nonvolatile) in memory
devices. "Circuitry", as used in any embodiment herein, may
comprise, for example, singly or in any combination, hardwired
circuitry, programmable circuitry such as computer processors
comprising one or more individual instruction processing cores,
state machine circuitry, and/or firmware that stores instructions
executed by programmable circuitry. The modules may, collectively
or individually, be embodied as circuitry that forms part of a
larger system, for example, an integrated circuit (IC), system
on-chip (SoC), desktop computers, laptop computers, tablet
computers, servers, smart phones, etc.
[0146] The terms and expressions which have been employed herein
are used as terms of description and not of limitation, and there
is no intention, in the use of such terms and expressions, of
excluding any equivalents of the features shown and described (or
portions thereof), and it is recognized that various modifications
are possible within the scope of the claims. Accordingly, the
claims are intended to cover all such equivalents.
* * * * *