Ubiquitous Inference of Mobility State of Human Custodian in - - PowerPoint PPT Presentation

▶
ubiquitous inference of mobility state of human custodian
SMART_READER_LITE
LIVE PREVIEW

Ubiquitous Inference of Mobility State of Human Custodian in - - PowerPoint PPT Presentation

Mar 05, 2023 276 likes •639 views

Ubiquitous Inference of Mobility State of Human Custodian in People-Centric Context Sensing Mattia Gustarini, Katarzyna Wac Institute of Services Science Quality of Life Group FACULTY OF ECONOMIC AND SOCIAL SCIENCES Department of Management

slide-1
SLIDE 1

FACULTY OF ECONOMIC AND SOCIAL SCIENCES

Department of Management Studies

Ubiquitous Inference of Mobility State of Human Custodian in People-Centric Context Sensing

Mattia Gustarini, Katarzyna Wac Institute of Services Science Quality of Life Group

slide-2
SLIDE 2

FACULTY OF ECONOMIC AND SOCIAL SCIENCES

Department of Management Studies

  • Some people-centric sensing challenges
  • capture of person’s mobility
  • understanding of context changes
  • preservation of user privacy

Motivation

2

slide-3
SLIDE 3

FACULTY OF ECONOMIC AND SOCIAL SCIENCES

Department of Management Studies

Goal

  • Infer mobile-fixed context of the

human custodian

  • accurately and efficiently (battery)
  • enable dynamic changes of the sensors’

duty cycle length

3

slide-4
SLIDE 4

FACULTY OF ECONOMIC AND SOCIAL SCIENCES

Department of Management Studies

Mobility Sensor

Raw Data Collection

4

slide-5
SLIDE 5

FACULTY OF ECONOMIC AND SOCIAL SCIENCES

Department of Management Studies

Mobility Sensor

Raw Data Collection

4

slide-6
SLIDE 6

FACULTY OF ECONOMIC AND SOCIAL SCIENCES

Department of Management Studies

Mobility Sensor

Raw Data Collection

4

slide-7
SLIDE 7

FACULTY OF ECONOMIC AND SOCIAL SCIENCES

Department of Management Studies

Mobility Sensor

Cell ID RSSI (dBm) Raw Data Collection

4

slide-8
SLIDE 8

FACULTY OF ECONOMIC AND SOCIAL SCIENCES

Department of Management Studies

Mobility Sensor

Cell ID RSSI (dBm) 1 2 Sessions numbered consecutively from 1 to N Raw Data Collection

4

slide-9
SLIDE 9

FACULTY OF ECONOMIC AND SOCIAL SCIENCES

Department of Management Studies

Mobility Sensor

Cell ID RSSI (dBm) 2s 7 scans per session 1 2 Sessions numbered consecutively from 1 to N Raw Data Collection

4

slide-10
SLIDE 10

FACULTY OF ECONOMIC AND SOCIAL SCIENCES

Department of Management Studies

Mobility Sensor

Raw Data Collection

64567

  • 90
  • 95
  • 89
  • 86
  • 91
  • 90
  • 87

4

5

CellID Alive sessions Last session scans 65784

  • 75
  • 80
  • 72
  • 74

5 61254

  • 81
  • 86
  • 89

2

slide-11
SLIDE 11

FACULTY OF ECONOMIC AND SOCIAL SCIENCES

Department of Management Studies

Mobility Sensor

Derive 3 features

6

slide-12
SLIDE 12

FACULTY OF ECONOMIC AND SOCIAL SCIENCES

Department of Management Studies

Mobility Sensor

Derive 3 features

Features

6

slide-13
SLIDE 13

FACULTY OF ECONOMIC AND SOCIAL SCIENCES

Department of Management Studies

Mobility Sensor

Derive 3 features

Features

Median life time of cells

6

Alive sessions c.

slide-14
SLIDE 14

FACULTY OF ECONOMIC AND SOCIAL SCIENCES

Department of Management Studies

Mobility Sensor

Derive 3 features 7 scans

Features

Median life time of cells Average euclidean distance of signals

6

slide-15
SLIDE 15

FACULTY OF ECONOMIC AND SOCIAL SCIENCES

Department of Management Studies

Mobility Sensor

Derive 3 features 7 scans

Features

Median life time of cells Average euclidean distance of signals Average fast wavelet transform signal range

6

slide-16
SLIDE 16

FACULTY OF ECONOMIC AND SOCIAL SCIENCES

Department of Management Studies

Mobility Sensor

Derive 3 features 7 scans

Features

Median life time of cells Average euclidean distance of signals Average fast wavelet transform signal range

6

Fixed Mobile

+

  • +
  • +
slide-17
SLIDE 17

FACULTY OF ECONOMIC AND SOCIAL SCIENCES

Department of Management Studies

Mobility Sensor

Tree Classifier

7

slide-18
SLIDE 18

FACULTY OF ECONOMIC AND SOCIAL SCIENCES

Department of Management Studies

Mobility Sensor

Tree Classifier 3 features

7

slide-19
SLIDE 19

FACULTY OF ECONOMIC AND SOCIAL SCIENCES

Department of Management Studies

Mobility Sensor

Tree Classifier 3 features

Tree Classifier

7

slide-20
SLIDE 20

FACULTY OF ECONOMIC AND SOCIAL SCIENCES

Department of Management Studies

Mobility Sensor

Tree Classifier 3 features

Tree Classifier

7

FIXED MOBILE

  • r
slide-21
SLIDE 21

FACULTY OF ECONOMIC AND SOCIAL SCIENCES

Department of Management Studies

Preliminary Experiments

  • Android phone
  • 1 user, 5 days, 1 phone operator
  • Mobility Sensor vs. accelerometer, network

location and GPS

  • mobile and fixed states predictions
  • battery consumption
  • User labeled the data (ESM with widget)

8

slide-22
SLIDE 22

FACULTY OF ECONOMIC AND SOCIAL SCIENCES

Department of Management Studies

Results

  • 539 predictions
  • 52% Fixed
  • 48% Mobile
  • 750 battery measurements

9

slide-23
SLIDE 23

FACULTY OF ECONOMIC AND SOCIAL SCIENCES

Department of Management Studies

Results

10

slide-24
SLIDE 24

FACULTY OF ECONOMIC AND SOCIAL SCIENCES

Department of Management Studies

Results

10

slide-25
SLIDE 25

FACULTY OF ECONOMIC AND SOCIAL SCIENCES

Department of Management Studies

Results

10

slide-26
SLIDE 26

FACULTY OF ECONOMIC AND SOCIAL SCIENCES

Department of Management Studies

Results

10

slide-27
SLIDE 27

FACULTY OF ECONOMIC AND SOCIAL SCIENCES

Department of Management Studies

Results

10

slide-28
SLIDE 28

FACULTY OF ECONOMIC AND SOCIAL SCIENCES

Department of Management Studies

Results

10

slide-29
SLIDE 29

FACULTY OF ECONOMIC AND SOCIAL SCIENCES

Department of Management Studies

Results

10

slide-30
SLIDE 30

FACULTY OF ECONOMIC AND SOCIAL SCIENCES

Department of Management Studies

Results

correct wrong n/a Accuracy Mobility methods

GPS Mobility Sensor Network Accelerometer 20 40 60 80 100

10

slide-31
SLIDE 31

FACULTY OF ECONOMIC AND SOCIAL SCIENCES

Department of Management Studies

Results

correct wrong n/a Accuracy Mobility methods

GPS Mobility Sensor Network Accelerometer 20 40 60 80 100

10

slide-32
SLIDE 32

FACULTY OF ECONOMIC AND SOCIAL SCIENCES

Department of Management Studies

Results

fix mobile Confusion Mobility methods

GPS Mobility Sensor Network Accelerometer 20 40 60 80 100

correct wrong n/a Accuracy Mobility methods

GPS Mobility Sensor Network Accelerometer 20 40 60 80 100

10

slide-33
SLIDE 33

FACULTY OF ECONOMIC AND SOCIAL SCIENCES

Department of Management Studies

Identified Problems

  • Network coverage
  • When fixed, network cell ping / pong
  • When mobile, minimum number of cells

11

slide-34
SLIDE 34

FACULTY OF ECONOMIC AND SOCIAL SCIENCES

Department of Management Studies

Ongoing Work

  • Improve the algorithm
  • Large case study involving real users
  • Mobile phone heterogeneity
  • neighbor CellIDs not always available
  • hardware battery consumption details
  • Experience Sampling Method

12

slide-35
SLIDE 35

FACULTY OF ECONOMIC AND SOCIAL SCIENCES

Department of Management Studies

Thank you!

Mattia Gustarini mattia.gustarini@unige.ch Katarzyna Wac katarzyna.wac@unige.ch http://www.qol.unige.ch

13