Transcript Reconciliation of regional travel model and passive device

Reconciliation of regional travel model
and passive device tracking data
14th TRB Planning Applications Conference
Leta F. Huntsinger
Rick Donnelly
Introduction
Passively collected mobile phone data has shown promise
as a low cost option for obtaining travel data:





Speed data (Using Cell Phone Technology to Collect Travel
Data, Kyle Ward)
Trip tables (Origin Destination Study using Cellular Technology
for Mobile, Al, Kevin Harrison)
Freight Data (Freight Data Collection Technique and Algorithm
using Cellular Phone and GIS Data, Ming-Heng Wang, et. al.)
other
Comparison of passively collected data against
traditionally collected survey data

2
Challenges
Household surveys





behaviorally rich, but small sample size at TAZ to TAZ level
Small TAZ to TAZ observations limit our understanding of
flows at the sub-district level
Many small MPOs cannot afford household surveys
Trip distribution parameters are the most challenging to
transfer
Passively collected data


3
Large sample size, but lacks behavioral richness
Data – Air Sage
4
Triangle Regional Model
5
Process
Disagg to TAZ
Convert AirSage
person trips to
vehicle trips
AM peak hour
assignment of
AirSage
Apply factors to
AirSage matrix
Apply AM
factors to
AirSage matrix
AM peak hour
assignment of
TRM
Add IE, EI, and
EE trips to
AirSage matrix
Develop AM
factors from
TRM data
Summarize
MOEs and
compare
6
Results – travel time comparisons
Trip Length Distribution (CongTT mins)
7.00
6.00
5.00
4.00
TRM – slightly
higher % of shorter
trips
3.00
2.00
1.00
1
5
9
13
17
21
25
29
33
37
41
45
49
53
57
61
65
69
73
77
81
85
89
93
97
101
105
109
113
117
121
0.00
TRM Percent
7
AirSage Percent
Average Trip Length (TT)
TRM
14.42
Air Sage
15.51
Results – district to district flows
District Map
District Trip Table Color Coded by Absolute and Relative Error
1
1
2
3
4
5
6
7
8
9
10
11
12
8
2
3
4
5
6
7
8
9
10
11
12
Results – Assignment MOEs
Functional Classification 23 – 26 are rural facilities
9
Results – Assignment MOEs
10
Results – Assignment MOEs
11
Results – Assignment MOEs
12
Findings and Recommendations








Early data set – includes Sprint data only
Great source of validation data
Low cost option
Lacks behavioral richness of household survey
Larger sample than household survey
Continuing improvements are needed
Useful to validate an estimated trip table
Add to toolbox
13
Acknowledgements



Co-author – Rick Donnelly
Kyle Ward, CAMPO
Air Sage
14
15