Indoor Localization with a Crowdsourcing based Fingerprints Collecting

ticketdonkeyΤεχνίτη Νοημοσύνη και Ρομποτική

25 Νοε 2013 (πριν από 3 χρόνια και 10 μήνες)

83 εμφανίσεις


Copyright ©
2013
by
SJTU, IWCT.

Dongchuan Road #800, Minhang,

Shanghai,200240

All rights reserved.

Indoor Localization with a Crowdsourcing
based Fingerprints Collecting


Copyright ©
2013
by
SJTU, IWCT.

Dongchuan Road #800, Minhang,

Shanghai,200240

All rights reserved.

System Architecture

Kernel Density Estimate
Sufficient Statistics
Extract Fingerprint
Optimum Reception Theory
Clustering
Affinity Propagation
Crowdsourced Process
User A
Device A
User B
Device B
User C
Device C
Crowdsourced Fingerprint Collection
Location Algorithm
K Nearest Neighbor
(
KNN
)
User Upload Rss Value
Use Any Device
User A
Device A
Cluster Matching
Affinity Propagation
Grid Window Filter
Restrict Estimate Results into
Sub Regions
AP Detection
Remove Aps below
threshold
Get Estimate Location
Information
Cloud Computing Platform
-

CloudFoundry
Location Process Using Fingerprint Database
Cloud Computing Platform
-

CloudFoundry
MMC
-
KNN
FingerPrint Database
for Diverse Devices
FingerPrint Database
for Diverse Devices

Copyright ©
2013
by
SJTU, IWCT.

Dongchuan Road #800, Minhang,

Shanghai,200240

All rights reserved.


Crowdsourcing based fingerprint extraction methods




Localization Algorithms based on clustering theory

Key Technology


Copyright ©
2013
by
SJTU, IWCT.

Dongchuan Road #800, Minhang,

Shanghai,200240

All rights reserved.


In crowdsourcing model, multiple users will upload
fingerprints via diverse devices


Our method extract fingerprint value based on RSS
probability estimation, choose the optimum value
from upload samples


Kernel density estimation eliminates device diversity
than Gaussian probability estimation


Fingerprints Extraction


Copyright ©
2013
by
SJTU, IWCT.

Dongchuan Road #800, Minhang,

Shanghai,200240

All rights reserved.


Comparison of Gaussian and Kernel density
estimation:


Fingerprints Extraction


Copyright ©
2013
by
SJTU, IWCT.

Dongchuan Road #800, Minhang,

Shanghai,200240

All rights reserved.


Based on kernel density estimation, choose optimum
value from multiple upload RSS samples by multiple
users by diverse devices.

Fingerprints Extraction


Copyright ©
2013
by
SJTU, IWCT.

Dongchuan Road #800, Minhang,

Shanghai,200240

All rights reserved.


MMC
-
KNN algorithm: find M most matched clusters,
then apply KNN principle to choose out matched
fingerprint


Use affinity propagation to process clustering:


Localization Algorithm: MMC
-
KNN


Copyright ©
2013
by
SJTU, IWCT.

Dongchuan Road #800, Minhang,

Shanghai,200240

All rights reserved.


How to find out the M most matched cluster?


Consider uploaded observation’s connections and
similarities with all exemplars


Apply affinity propagation again and get
responsibility vector:



choose the M most matched cluster by sort this
responsibility vector



Localization Algorithm: MMC
-
KNN


Copyright ©
2013
by
SJTU, IWCT.

Dongchuan Road #800, Minhang,

Shanghai,200240

All rights reserved.


Assign a weight factor to each cluster’s fingerprints




Apply a grid window filter to filter a region which
has the maximum weight, with the purpose to restrict
KNN applied to a bursting region

Localization Algorithm: MMC
-
KNN

( )
(,)
c
e
w f
D f o


Copyright ©
2013
by
SJTU, IWCT.

Dongchuan Road #800, Minhang,

Shanghai,200240

All rights reserved.


Average error distance with different matched cluster
number and grid window size for Nexus
-
S

Real
-
time experimental
testbed


Copyright ©
2013
by
SJTU, IWCT.

Dongchuan Road #800, Minhang,

Shanghai,200240

All rights reserved.


220 observation’s error distance statistic with best
performance parameters for Nexus
-
S


Real
-
time experimental
testbed


Copyright ©
2013
by
SJTU, IWCT.

Dongchuan Road #800, Minhang,

Shanghai,200240

All rights reserved.


CDF of location error distance for different
algorithms


Real
-
time experimental
testbed


Copyright ©
2013
by
SJTU, IWCT.

Dongchuan Road #800, Minhang,

Shanghai,200240

All rights reserved.


Comparison of different types devices’ location
performance under diverse fingerprint databases

Real
-
time experimental
testbed