Automated Traffic Signal Performance Measures PDFS/3C... · All of our metrics are based on ....
Transcript of Automated Traffic Signal Performance Measures PDFS/3C... · All of our metrics are based on ....
![Page 1: Automated Traffic Signal Performance Measures PDFS/3C... · All of our metrics are based on . outputs. not Objectives . Source: INDOT/FHWA. One “typical” day. Count a few peak](https://reader033.fdocuments.in/reader033/viewer/2022060520/604e8d44a759c143251af79c/html5/thumbnails/1.jpg)
Penn State Transportation Engineering & Safety ConferenceDecember 6, 2018
Automated Traffic Signal Performance Measures
![Page 2: Automated Traffic Signal Performance Measures PDFS/3C... · All of our metrics are based on . outputs. not Objectives . Source: INDOT/FHWA. One “typical” day. Count a few peak](https://reader033.fdocuments.in/reader033/viewer/2022060520/604e8d44a759c143251af79c/html5/thumbnails/2.jpg)
2
Acceptable signal operation?
![Page 3: Automated Traffic Signal Performance Measures PDFS/3C... · All of our metrics are based on . outputs. not Objectives . Source: INDOT/FHWA. One “typical” day. Count a few peak](https://reader033.fdocuments.in/reader033/viewer/2022060520/604e8d44a759c143251af79c/html5/thumbnails/3.jpg)
3
Signal Timing Development in Pennsylvania
“The way we’ve always done it”
LOS
![Page 4: Automated Traffic Signal Performance Measures PDFS/3C... · All of our metrics are based on . outputs. not Objectives . Source: INDOT/FHWA. One “typical” day. Count a few peak](https://reader033.fdocuments.in/reader033/viewer/2022060520/604e8d44a759c143251af79c/html5/thumbnails/4.jpg)
4
Is this better?
LOS
![Page 5: Automated Traffic Signal Performance Measures PDFS/3C... · All of our metrics are based on . outputs. not Objectives . Source: INDOT/FHWA. One “typical” day. Count a few peak](https://reader033.fdocuments.in/reader033/viewer/2022060520/604e8d44a759c143251af79c/html5/thumbnails/5.jpg)
5
Traditional: Model a lot… Measure a little
S M T W T F S
1 2 3 4 5 6 7
8 9 10 11 12 13 14
15 16 17 18 19 20 21
22 23 24 25 26 27 28
29 30
APRIL
All of our metrics are based on outputs not Objectives
Source: INDOT/FHWA
One “typical” day Count a few peak periods
Find peak hour
Apply peak hour factorAnalyze peak 15 min.
2 peak hours on 1 day =0.005% of a year
Outputs:Cycle Length
SplitsOffsets
![Page 6: Automated Traffic Signal Performance Measures PDFS/3C... · All of our metrics are based on . outputs. not Objectives . Source: INDOT/FHWA. One “typical” day. Count a few peak](https://reader033.fdocuments.in/reader033/viewer/2022060520/604e8d44a759c143251af79c/html5/thumbnails/6.jpg)
6
Is this better?
Data Collection
Modeling
![Page 7: Automated Traffic Signal Performance Measures PDFS/3C... · All of our metrics are based on . outputs. not Objectives . Source: INDOT/FHWA. One “typical” day. Count a few peak](https://reader033.fdocuments.in/reader033/viewer/2022060520/604e8d44a759c143251af79c/html5/thumbnails/7.jpg)
7
A Better Way
Trigger Collect Data
Model & Optimize
Implement & Fine Tune
ComplaintsDeveloper Impact
Outputs:Cycle Length
SplitsOffsets
Review MetricsField Observation
Model Optimize Implement & Fine Tune
Output:Offsets
Traditional Process (Reactive)
Modified Process with Performance Data (Proactive)
How do measures compare to
agency goals?
Adapted from UDOT
![Page 8: Automated Traffic Signal Performance Measures PDFS/3C... · All of our metrics are based on . outputs. not Objectives . Source: INDOT/FHWA. One “typical” day. Count a few peak](https://reader033.fdocuments.in/reader033/viewer/2022060520/604e8d44a759c143251af79c/html5/thumbnails/8.jpg)
8
Using Data to Measure Arterial Performance
• High resolution data• UDOT ATSPM software
Corridor Level Intersection Level
• Probe speed data• RITIS/PDA Suite
![Page 9: Automated Traffic Signal Performance Measures PDFS/3C... · All of our metrics are based on . outputs. not Objectives . Source: INDOT/FHWA. One “typical” day. Count a few peak](https://reader033.fdocuments.in/reader033/viewer/2022060520/604e8d44a759c143251af79c/html5/thumbnails/9.jpg)
Corridor Level Metrics
Phase 1: Proof of ConceptPhase 2: Enabling Access, Scalability, and Usability
![Page 10: Automated Traffic Signal Performance Measures PDFS/3C... · All of our metrics are based on . outputs. not Objectives . Source: INDOT/FHWA. One “typical” day. Count a few peak](https://reader033.fdocuments.in/reader033/viewer/2022060520/604e8d44a759c143251af79c/html5/thumbnails/10.jpg)
10
Probe Data
• Pennsylvania INRIX coverage– 25,000 TMC segments = 16,600 miles– 112,000 XD segments = 23,200 miles (20,200 on arterials)
![Page 11: Automated Traffic Signal Performance Measures PDFS/3C... · All of our metrics are based on . outputs. not Objectives . Source: INDOT/FHWA. One “typical” day. Count a few peak](https://reader033.fdocuments.in/reader033/viewer/2022060520/604e8d44a759c143251af79c/html5/thumbnails/11.jpg)
11
Research / Proof of Concept
• TRB Annual Meeting 2017 Paper # 17-00314– http://docs.trb.org/prp/17-00314.pdf
• Proof of Concept– 138 “Super-Critical” corridors in Philadelphia area– Covered 2,184 signals on 766 miles of arterials
![Page 12: Automated Traffic Signal Performance Measures PDFS/3C... · All of our metrics are based on . outputs. not Objectives . Source: INDOT/FHWA. One “typical” day. Count a few peak](https://reader033.fdocuments.in/reader033/viewer/2022060520/604e8d44a759c143251af79c/html5/thumbnails/12.jpg)
12
Travel Time Comparison
• Travel Time• Reliability
0%
25%
50%
75%
100%
10 11 12 13 14 15 16 17 18
Cum
ulat
ive
Freq
uenc
y
Travel Time (min.)
Improved Travel TimeImproved Reliability
Before
After
Retiming Week
S M T W T F S
1 2 3
4 5 6 7 8 9 10
11 12 13 14 15 16 17
18 19 20 21 22 23 24
25 26 27 28 29 30 31
MARCH 2012
S M T W T F S
1 2 3 4 5 6 7
8 9 10 11 12 13 14
15 16 17 18 19 20 21
22 23 24 25 26 27 28
29 30
APRIL 2012
![Page 13: Automated Traffic Signal Performance Measures PDFS/3C... · All of our metrics are based on . outputs. not Objectives . Source: INDOT/FHWA. One “typical” day. Count a few peak](https://reader033.fdocuments.in/reader033/viewer/2022060520/604e8d44a759c143251af79c/html5/thumbnails/13.jpg)
13
Travel Time Normalization
• Normalize corridors of difference length and speed limits• Identify corridors with below average performance
Normalize travel times
Median TTSpeed Limit TT
![Page 14: Automated Traffic Signal Performance Measures PDFS/3C... · All of our metrics are based on . outputs. not Objectives . Source: INDOT/FHWA. One “typical” day. Count a few peak](https://reader033.fdocuments.in/reader033/viewer/2022060520/604e8d44a759c143251af79c/html5/thumbnails/14.jpg)
14
Travel Time Normalization
![Page 15: Automated Traffic Signal Performance Measures PDFS/3C... · All of our metrics are based on . outputs. not Objectives . Source: INDOT/FHWA. One “typical” day. Count a few peak](https://reader033.fdocuments.in/reader033/viewer/2022060520/604e8d44a759c143251af79c/html5/thumbnails/15.jpg)
15
Interquartile-Range (IQR) Normalization
0%
25%
50%
75%
100%
10 11 12 13 14 15 16 17 18
Cum
ulat
ive
Freq
uenc
y 11.8 mins
75% TT
25% TT
13 mins
Normalized IQR = (75𝑡𝑡𝑡 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑡𝑡𝑝𝑝𝑝𝑝𝑝𝑝 𝑇𝑇𝑇𝑇−25𝑡𝑡𝑡 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑡𝑡𝑝𝑝𝑝𝑝𝑝𝑝 𝑇𝑇𝑇𝑇)𝑆𝑆𝑝𝑝𝑝𝑝𝑝𝑝𝑆𝑆 𝑝𝑝𝑝𝑝𝑙𝑙𝑝𝑝𝑡𝑡 𝑇𝑇𝑇𝑇
![Page 16: Automated Traffic Signal Performance Measures PDFS/3C... · All of our metrics are based on . outputs. not Objectives . Source: INDOT/FHWA. One “typical” day. Count a few peak](https://reader033.fdocuments.in/reader033/viewer/2022060520/604e8d44a759c143251af79c/html5/thumbnails/16.jpg)
16
Travel Time Delta Ranking
Reliability vs Central Tendency
Interquartile-Range (IQR) Normalization RankingTravel Time Normalization Ranking
![Page 17: Automated Traffic Signal Performance Measures PDFS/3C... · All of our metrics are based on . outputs. not Objectives . Source: INDOT/FHWA. One “typical” day. Count a few peak](https://reader033.fdocuments.in/reader033/viewer/2022060520/604e8d44a759c143251af79c/html5/thumbnails/17.jpg)
17
PDA Suite: Corridor and Time Selection
![Page 18: Automated Traffic Signal Performance Measures PDFS/3C... · All of our metrics are based on . outputs. not Objectives . Source: INDOT/FHWA. One “typical” day. Count a few peak](https://reader033.fdocuments.in/reader033/viewer/2022060520/604e8d44a759c143251af79c/html5/thumbnails/18.jpg)
18
PDA Suite: Travel Time Comparison
![Page 19: Automated Traffic Signal Performance Measures PDFS/3C... · All of our metrics are based on . outputs. not Objectives . Source: INDOT/FHWA. One “typical” day. Count a few peak](https://reader033.fdocuments.in/reader033/viewer/2022060520/604e8d44a759c143251af79c/html5/thumbnails/19.jpg)
19
PDA Suite: Cumulative Distribution Chart
![Page 20: Automated Traffic Signal Performance Measures PDFS/3C... · All of our metrics are based on . outputs. not Objectives . Source: INDOT/FHWA. One “typical” day. Count a few peak](https://reader033.fdocuments.in/reader033/viewer/2022060520/604e8d44a759c143251af79c/html5/thumbnails/20.jpg)
20
PDA Suite: Change Between Dates
![Page 21: Automated Traffic Signal Performance Measures PDFS/3C... · All of our metrics are based on . outputs. not Objectives . Source: INDOT/FHWA. One “typical” day. Count a few peak](https://reader033.fdocuments.in/reader033/viewer/2022060520/604e8d44a759c143251af79c/html5/thumbnails/21.jpg)
21
PDA Suite: Change Between Dates
Significant improvements to
fastest travel times on the
corridor
Mixed results for median travel
along the corridor
![Page 22: Automated Traffic Signal Performance Measures PDFS/3C... · All of our metrics are based on . outputs. not Objectives . Source: INDOT/FHWA. One “typical” day. Count a few peak](https://reader033.fdocuments.in/reader033/viewer/2022060520/604e8d44a759c143251af79c/html5/thumbnails/22.jpg)
22
PDA Suite: Change Between Dates
Central portion of corridor got worse in westbound direction
![Page 23: Automated Traffic Signal Performance Measures PDFS/3C... · All of our metrics are based on . outputs. not Objectives . Source: INDOT/FHWA. One “typical” day. Count a few peak](https://reader033.fdocuments.in/reader033/viewer/2022060520/604e8d44a759c143251af79c/html5/thumbnails/23.jpg)
23
PDA Suite: Travel Time Delta Ranking
![Page 24: Automated Traffic Signal Performance Measures PDFS/3C... · All of our metrics are based on . outputs. not Objectives . Source: INDOT/FHWA. One “typical” day. Count a few peak](https://reader033.fdocuments.in/reader033/viewer/2022060520/604e8d44a759c143251af79c/html5/thumbnails/24.jpg)
24
PDA Suite: Travel Time Delta Ranking (Slope Chart)
Median and IQR both got worse
Median and IQR both got better
![Page 25: Automated Traffic Signal Performance Measures PDFS/3C... · All of our metrics are based on . outputs. not Objectives . Source: INDOT/FHWA. One “typical” day. Count a few peak](https://reader033.fdocuments.in/reader033/viewer/2022060520/604e8d44a759c143251af79c/html5/thumbnails/25.jpg)
Intersection-Level Metrics
![Page 26: Automated Traffic Signal Performance Measures PDFS/3C... · All of our metrics are based on . outputs. not Objectives . Source: INDOT/FHWA. One “typical” day. Count a few peak](https://reader033.fdocuments.in/reader033/viewer/2022060520/604e8d44a759c143251af79c/html5/thumbnails/26.jpg)
26
High Res Data: A Fitness Tracker for Traffic Signals
Source: FHWA
Data Analysis &Performance Reporting
High ResolutionData Collection
![Page 27: Automated Traffic Signal Performance Measures PDFS/3C... · All of our metrics are based on . outputs. not Objectives . Source: INDOT/FHWA. One “typical” day. Count a few peak](https://reader033.fdocuments.in/reader033/viewer/2022060520/604e8d44a759c143251af79c/html5/thumbnails/27.jpg)
27
ATSPM Implementation
Controller Detection Communication
![Page 28: Automated Traffic Signal Performance Measures PDFS/3C... · All of our metrics are based on . outputs. not Objectives . Source: INDOT/FHWA. One “typical” day. Count a few peak](https://reader033.fdocuments.in/reader033/viewer/2022060520/604e8d44a759c143251af79c/html5/thumbnails/28.jpg)
28
Implementing ATSPM: Controllers
• Buy a new controller for about the same price as doing one round of count, analyze & retime
• Vendor neutral – Bulletin 15 #’s in bold below– Links to manufacturer product sheets below
Econolite Cobalt: Any version ECO-127P, 128P & 129PEconolite ASC3 NEMA: v2.50+ & OS 1.14.03+ ECO-116PEconolite 2070 w/ 1C CPU: v 32.50+ Peek ATC: Greenwave 03.05.0528+ PTS-042P
McCain ATC Omni eX 1.6+ MCC-018P
Trafficware Model 980ATC: v 76.10+ TFW-011PTrafficware ATC Controller: v76.10+ Intelight: Maxtime v1.7.0+ INT-009P, 010P & 011P
Siemens M50 LinuxSiemens M60 ATC SMS-231PECOM v 3.52+NTCIP v 4.53+
![Page 29: Automated Traffic Signal Performance Measures PDFS/3C... · All of our metrics are based on . outputs. not Objectives . Source: INDOT/FHWA. One “typical” day. Count a few peak](https://reader033.fdocuments.in/reader033/viewer/2022060520/604e8d44a759c143251af79c/html5/thumbnails/29.jpg)
29
Implementing ATSPM: Detection
![Page 30: Automated Traffic Signal Performance Measures PDFS/3C... · All of our metrics are based on . outputs. not Objectives . Source: INDOT/FHWA. One “typical” day. Count a few peak](https://reader033.fdocuments.in/reader033/viewer/2022060520/604e8d44a759c143251af79c/html5/thumbnails/30.jpg)
30
Implementing ATSPM: Communication
Strike-off Letter (SOL) 494-16-02(Effective: February 12, 2016)
![Page 31: Automated Traffic Signal Performance Measures PDFS/3C... · All of our metrics are based on . outputs. not Objectives . Source: INDOT/FHWA. One “typical” day. Count a few peak](https://reader033.fdocuments.in/reader033/viewer/2022060520/604e8d44a759c143251af79c/html5/thumbnails/31.jpg)
31
Implementing ATSPM: No Communication
![Page 32: Automated Traffic Signal Performance Measures PDFS/3C... · All of our metrics are based on . outputs. not Objectives . Source: INDOT/FHWA. One “typical” day. Count a few peak](https://reader033.fdocuments.in/reader033/viewer/2022060520/604e8d44a759c143251af79c/html5/thumbnails/32.jpg)
32
ATSPM: Improve progression on critical corridors
• Purdue Coordination Diagram
Few arrivals on red
%AoG > % GT
![Page 33: Automated Traffic Signal Performance Measures PDFS/3C... · All of our metrics are based on . outputs. not Objectives . Source: INDOT/FHWA. One “typical” day. Count a few peak](https://reader033.fdocuments.in/reader033/viewer/2022060520/604e8d44a759c143251af79c/html5/thumbnails/33.jpg)
33
ATSPM: Split Failures
• Green Occupancy Ratio (GOR)– % of time detection zone occupied while signal is green for the
lane(s) served by the phase• Red Occupancy Ratio (ROR)
– % of time detection zone occupied during first 5 seconds of red for the lane(s) served by the phase
• Split Failure80%+ GOR
AND80%+ ROR
![Page 34: Automated Traffic Signal Performance Measures PDFS/3C... · All of our metrics are based on . outputs. not Objectives . Source: INDOT/FHWA. One “typical” day. Count a few peak](https://reader033.fdocuments.in/reader033/viewer/2022060520/604e8d44a759c143251af79c/html5/thumbnails/34.jpg)
34
ATSPM: Minimize delay for intersecting users
Split failures higher mid-morning and early afternoon than during
lunch hour
![Page 35: Automated Traffic Signal Performance Measures PDFS/3C... · All of our metrics are based on . outputs. not Objectives . Source: INDOT/FHWA. One “typical” day. Count a few peak](https://reader033.fdocuments.in/reader033/viewer/2022060520/604e8d44a759c143251af79c/html5/thumbnails/35.jpg)
35
Minimize delay for intersecting users
8 sec. programmed split in Plan 2!
(3 second green)
![Page 36: Automated Traffic Signal Performance Measures PDFS/3C... · All of our metrics are based on . outputs. not Objectives . Source: INDOT/FHWA. One “typical” day. Count a few peak](https://reader033.fdocuments.in/reader033/viewer/2022060520/604e8d44a759c143251af79c/html5/thumbnails/36.jpg)
36
Minimize delay for intersecting users
Consider starting Plan 4 earlier than 7:15 instead
of running free?
![Page 37: Automated Traffic Signal Performance Measures PDFS/3C... · All of our metrics are based on . outputs. not Objectives . Source: INDOT/FHWA. One “typical” day. Count a few peak](https://reader033.fdocuments.in/reader033/viewer/2022060520/604e8d44a759c143251af79c/html5/thumbnails/37.jpg)
37
ATSPM: Minimize delay for intersecting users
254 second cycle = 4 minutes of ped delay!
Ped recall = broken button
![Page 38: Automated Traffic Signal Performance Measures PDFS/3C... · All of our metrics are based on . outputs. not Objectives . Source: INDOT/FHWA. One “typical” day. Count a few peak](https://reader033.fdocuments.in/reader033/viewer/2022060520/604e8d44a759c143251af79c/html5/thumbnails/38.jpg)
38
ATSPM: Maintain equipment
209 second split?254 second cycle?
Programming error!
![Page 39: Automated Traffic Signal Performance Measures PDFS/3C... · All of our metrics are based on . outputs. not Objectives . Source: INDOT/FHWA. One “typical” day. Count a few peak](https://reader033.fdocuments.in/reader033/viewer/2022060520/604e8d44a759c143251af79c/html5/thumbnails/39.jpg)
39
ATSPM: Maintain equipment
Utility contractor hit loop
52 sec max. per permit
![Page 40: Automated Traffic Signal Performance Measures PDFS/3C... · All of our metrics are based on . outputs. not Objectives . Source: INDOT/FHWA. One “typical” day. Count a few peak](https://reader033.fdocuments.in/reader033/viewer/2022060520/604e8d44a759c143251af79c/html5/thumbnails/40.jpg)
40
ATSPM: Maintain equipment
Same phase before loop
cut
![Page 41: Automated Traffic Signal Performance Measures PDFS/3C... · All of our metrics are based on . outputs. not Objectives . Source: INDOT/FHWA. One “typical” day. Count a few peak](https://reader033.fdocuments.in/reader033/viewer/2022060520/604e8d44a759c143251af79c/html5/thumbnails/41.jpg)
41
ATSPM: Maintain equipment
Revised permit to
reduce max time to 26 seconds
Min recall to get cars
to loop past stop bar
that was still working
![Page 42: Automated Traffic Signal Performance Measures PDFS/3C... · All of our metrics are based on . outputs. not Objectives . Source: INDOT/FHWA. One “typical” day. Count a few peak](https://reader033.fdocuments.in/reader033/viewer/2022060520/604e8d44a759c143251af79c/html5/thumbnails/42.jpg)
42
ATSPM: Example (Shippensburg, PA)WB Left: Few split failures
Low green occupancy
![Page 43: Automated Traffic Signal Performance Measures PDFS/3C... · All of our metrics are based on . outputs. not Objectives . Source: INDOT/FHWA. One “typical” day. Count a few peak](https://reader033.fdocuments.in/reader033/viewer/2022060520/604e8d44a759c143251af79c/html5/thumbnails/43.jpg)
43
ATSPM: Example (Shippensburg, PA)
EB Thru:Split failures
< 25%
![Page 44: Automated Traffic Signal Performance Measures PDFS/3C... · All of our metrics are based on . outputs. not Objectives . Source: INDOT/FHWA. One “typical” day. Count a few peak](https://reader033.fdocuments.in/reader033/viewer/2022060520/604e8d44a759c143251af79c/html5/thumbnails/44.jpg)
44
ATSPM: Example (Shippensburg, PA)
EB Left:More split
failures
![Page 45: Automated Traffic Signal Performance Measures PDFS/3C... · All of our metrics are based on . outputs. not Objectives . Source: INDOT/FHWA. One “typical” day. Count a few peak](https://reader033.fdocuments.in/reader033/viewer/2022060520/604e8d44a759c143251af79c/html5/thumbnails/45.jpg)
45
Automated Traffic Signal Performance Measures
Steve Gault, P.E., PTOEStatewide Traffic Signal Operations EngineerPennsylvania Department of Transportation