EasyTracker Automatic Transit Tracking, Mapping, and Arrival Time - - PowerPoint PPT Presentation

▶
easytracker
SMART_READER_LITE
LIVE PREVIEW

EasyTracker Automatic Transit Tracking, Mapping, and Arrival Time - - PowerPoint PPT Presentation

Dec 19, 2023 3.96k likes •4.98k views

EasyTracker Automatic Transit Tracking, Mapping, and Arrival Time Prediction Using Smartphones James Biagioni, Tomas Gerlich, Timothy Merrifield and Jakob Eriksson We love bus trackers! slide 2 Winter in Chicago slide 3 Our shuttle web

slide-1
SLIDE 1

James Biagioni, Tomas Gerlich, Timothy Merrifield and Jakob Eriksson Automatic Transit Tracking, Mapping, and Arrival Time Prediction Using Smartphones

EasyTracker

slide-2
SLIDE 2

slide

We love bus trackers!

2

slide-3
SLIDE 3

slide

Winter in Chicago

3

slide-4
SLIDE 4

slide 4

Our shuttle web (before)

slide-5
SLIDE 5

slide 4

Our shuttle web (before)

slide-6
SLIDE 6

slide

Our shuttle web (after)

5

slide-7
SLIDE 7

slide

One service for everyone

6

slide-8
SLIDE 8

slide

Our shuttle web

7

slide-9
SLIDE 9

slide

Our shuttle web

7

slide-10
SLIDE 10

slide

This paper in a nutshell

  • Automatic generation of
  • route shapes
  • stop locations
  • schedules
  • Online processing for
  • vehicle-to-route classification
  • arrival-time prediction

8

slide-11
SLIDE 11

slide

EasyTracker installation

9

slide-12
SLIDE 12

slide

EasyTracker installation

  • 1. Obtain smartphone

9

slide-13
SLIDE 13

slide

EasyTracker installation

  • 1. Obtain smartphone

9

  • 2. Install EasyTracker app
slide-14
SLIDE 14

slide

EasyTracker installation

  • 1. Obtain smartphone

9

  • 2. Install EasyTracker app
  • 3. Stick phone in bus
slide-15
SLIDE 15

slide

  • 4. Relax

10

slide-16
SLIDE 16

slide

System overview

11

GPS GPS

slide-17
SLIDE 17

slide

System overview

11

GPS GPS

slide-18
SLIDE 18

slide

Batch processing

12

slide-19
SLIDE 19

slide

Batch processing

12

slide-20
SLIDE 20

slide

Raw GPS traces

13

slide-21
SLIDE 21

slide

Route map

14

slide-22
SLIDE 22

slide

Raw GPS traces

15

slide-23
SLIDE 23

slide

Kernel Density Estimation

16

slide-24
SLIDE 24

slide

Kernel Density Estimation

16

slide-25
SLIDE 25

slide

Kernel Density Estimation

16

slide-26
SLIDE 26

slide

Kernel Density Estimation

16

slide-27
SLIDE 27

slide

K(x) = 1 √ 2πσ2 e− x2

2σ2

Kernel density estimation

17

ˆ f(x) = 1 n

n

X

i=1

K(x − xi)

slide-28
SLIDE 28

slide

2-D histogram

18

slide-29
SLIDE 29

slide

Trajectory density estimate

19

slide-30
SLIDE 30

slide

Thresholded image

20

slide-31
SLIDE 31

slide

  • Davies et al., 2006

Map extraction

21

slide-32
SLIDE 32

slide

Route extraction

  • Map match GPS traces
  • Viterbi-based map matching
  • based on Thiagarajan, et al. 2009
  • Extract common routes
  • edge subsequence matching
  • statistical test removes spurious results

22

slide-33
SLIDE 33

slide

Route extraction

23

slide-34
SLIDE 34

slide

Route extraction

23

slide-35
SLIDE 35

slide

Route extraction

23

slide-36
SLIDE 36

slide

Route extraction

24

slide-37
SLIDE 37

slide

Route extraction

24

slide-38
SLIDE 38

slide

Route extraction

24

slide-39
SLIDE 39

slide

Route extraction

24

slide-40
SLIDE 40

slide

Route extraction

24

slide-41
SLIDE 41

slide

Route extraction

24

slide-42
SLIDE 42

slide

Route extraction

24

slide-43
SLIDE 43

slide

Route extraction

24

slide-44
SLIDE 44

slide

Route extraction

24

slide-45
SLIDE 45

slide

Route extraction

24

slide-46
SLIDE 46

slide

Route extraction

24

slide-47
SLIDE 47

slide

Route extraction

24

slide-48
SLIDE 48

slide

Route extraction results

25

slide-49
SLIDE 49

slide

Welch’s t-Test

26

0.008 0.015 0.023 0.030 P-value

slide-50
SLIDE 50

slide

Routes separated

27

Real Routes Spurious Routes

slide-51
SLIDE 51

slide

Stop extraction

28

slide-52
SLIDE 52

slide

Route-labeled GPS traces

29

slide-53
SLIDE 53

slide

2-D histogram

30

slide-54
SLIDE 54

slide

Point density estimate

31

slide-55
SLIDE 55

slide

Thresholded binary image

32

slide-56
SLIDE 56

slide

Noise in binary image

33

slide-57
SLIDE 57

slide

Noise reduced binary image

34

slide-58
SLIDE 58

slide

Stop extraction

35

slide-59
SLIDE 59

slide

Stop extraction

36

slide-60
SLIDE 60

slide

Stop extraction performance

37

slide-61
SLIDE 61

slide

Schedule extraction

38

slide-62
SLIDE 62

slide

Bus stop arrival times

39

slide-63
SLIDE 63

slide

Bus stop arrival times

39

slide-64
SLIDE 64

slide

Bus stop arrival times

39

slide-65
SLIDE 65

slide

Bus stop arrival times

39

slide-66
SLIDE 66

slide

Bus stop arrival times

39

slide-67
SLIDE 67

slide

Bus stop arrival times

40

slide-68
SLIDE 68

slide

First stop schedule

41

slide-69
SLIDE 69

slide

Travel time variance

42

slide-70
SLIDE 70

slide

Last stop arrival times

43

slide-71
SLIDE 71

slide

travel time(1, j, t) = 1 |D| X

D

at

j − at 1

Compute mean travel times

44

slide-72
SLIDE 72

slide

Compute downstream schedules

45

kj

t = k1 t + travel time(1, j, t)

slide-73
SLIDE 73

slide

Last stop arrival times

46

slide-74
SLIDE 74

slide

Last stop schedule

47

slide-75
SLIDE 75

slide

Schedule accuracy

48

slide-76
SLIDE 76

slide

Schedule accuracy

48

slide-77
SLIDE 77

slide

System architecture

49

slide-78
SLIDE 78

slide

Online processing

50

slide-79
SLIDE 79

slide

Online processing

50

slide-80
SLIDE 80

slide

Un-classified buses

51

slide-81
SLIDE 81

slide 52

Classified buses

slide-82
SLIDE 82

slide

Hidden Markov model

53

slide-83
SLIDE 83

slide

Classification accuracy

54 0.2 0.4 0.6 0.8 1.0 Correct Incorrect Unclassified

slide-84
SLIDE 84

slide

Classification delay

55

slide-85
SLIDE 85

slide

Classification delay

55

slide-86
SLIDE 86

slide

Arrival time prediction

56

slide-87
SLIDE 87

slide

Predicting arrival times

57

time until arrival(si) = γtravel time(sprev+1, si)+ (1 − γ)travel time(sprev, si)

slide-88
SLIDE 88

slide

Arrival time predictions

58

slide-89
SLIDE 89

slide

Schedule vs. real-time

59

slide-90
SLIDE 90

slide

Schedule vs. real-time

59

slide-91
SLIDE 91

slide

Schedule vs. real-time

59

slide-92
SLIDE 92

slide

System overview

60

GPS GPS

slide-93
SLIDE 93

slide

Come and see our demo!

  • Thursday, 3:30p-7:30p

61

slide-94
SLIDE 94

Questions?

Thanks!