Adaptive Traffic Signal Control WVDOH/MPO/FHWA Planning Conference 10/3/2012 Andrew P. Nichols, PhD, PE

Adaptive Traffic Signal Control
WVDOH/MPO/FHWA Planning Conference
10/3/2012
Andrew P. Nichols, PhD, PE
Rahall Transportation Institute
Presentation Overview

Traffic Signals 101





What is Adaptive Signal Control?
Adaptive Systems in WV
Preliminary Evaluation Results



Types of Traffic Signal Control
Terminology Overview
Simulation
Field Data
Planning for Traffic Signals
Types of Traffic Signal Control

Basic (Free) Operation




Coordinated (Time-of-Day) Operation




With or without vehicle/pedestrian detection
Commonly used for isolated intersections
Green times programmed based on anticipated demand
With or without vehicle/pedestrian detection
Used in corridor or downtown grid signal systems
Cycle, Offset, and Splits are the key parameters that are typically
programmed for anticipated demand at different time periods
Adaptive Operation


Requires vehicle detection
Most systems dynamically adjust Cycle, Offset, and Splits based on
current or historical demand
What is a Phase Split?


Each signalized intersection movement is a phase
4-way intersection with left-turn arrows has 8 phases
Split Time = Green + Yellow + All Red
Movement
Phase
66 1
2
3
4
5
6
7
8
7 44
1
3 88

5 22
What is a Phase Split?


Phase split times are programmed to account for demand
Duration can also be shortened with vehicle detection
1
5
2
3
6
7
4
8
Phase 8 Split = G + Y + R
What is a Cycle?

The total time it takes to serve all phase splits
Cycle = Split 1 + Split 2 + Split 3 + Split 4
Cycle = Split 5 + Split 6 + Split 7 + Split 8
1
5
2
3
6
7
4
8
What is an Offset?

Programmed for the “mainline” movement at each
intersection to provide progression through the system


1
i.e., Offsetting the start of green
Consider a Northbound Phase 2 movement only:
2
3
4
1
5
6
7
8
2
3
4
Time-Space Diagram (NB)
2
1
2
3
4
1
2
3
4
1
1
2
3
4
1
2
3
4
1
1
2
3
4
1
2
3
4
1
Space (feet)
3
1
8:00:00
Time (seconds)
Time-Space Diagram (Zero Offset)
2
1
Vehicle
Trajectory
Space (feet)
3
2
2
2
2
2
2
8:00:00
Time (seconds)
Time-Space Diagram (w/ offsets)
3
Space (feet)
2
2
2
Offset 2
2
2
Offset 1
1
Vehicle
Trajectory
2
2
8:00:00
Time (seconds)
Time-Space Diagram (w/ offsets)
3
2
Offset 2
Space (feet)
2
2
2
2
Offset 1
1
2
2
8:00:00
Time (seconds)
Coordination Plans


Traffic engineers can deploy multiple coordination plans
to accommodate different traffic patterns
How are these plans derived?
Traffic Data Collection

Optimization Software
Field Tuning
This process cannot account for all traffic patterns!
Star City Bridge, Morgantown
Inbound Traffic on Saturday

When to start each Coordination Plan?
Saturday 1
Plan 1 Plan 2
Saturday 2
Plan 3
Plan 1 Plan 2
Plan 3
Oops!
Star City Bridge, Morgantown
Saturday Traffic
Saturday 1

Should offsets
provide progression
for INBOUND or
OUTBOUND
traffic?
Inbound
Outbound
Favor
Outbound!
Saturday 2
Oops!
Inbound
Outbound
What is Adaptive Signal Control?



Respond more intelligently to fluctuations in traffic
patterns (beyond shortening phase times with vehicle
detection)
All adaptive systems require vehicle detection
Adjust phase split times


Adjust cycle length


More/less time for a left-turn movements or side streets
Longer for congested periods, shorter for off-peak
Adjust offsets


Accommodate inbound vs. outbound traffic
Account for change in traffic speeds due to severe weather
Adaptive Signal Control

20+ adaptive control systems on the market
<1% of signals nationwide use adaptive technology
NONE of the systems are “plug and play”
Each system is unique

FHWA Every Day Counts Initiative






Identified ASC as being significantly underutilized
Providing some guidance on how to select a system
Developing guidelines on how to evaluate performance
Adaptive Systems in WV

WVDOH operates traffic signals statewide



Adaptive Signal Control Deployments






~1,900 traffic signals
~107 closed loop signal systems (running coordinated timings)
(16) Morgantown WV-705 Corridor (In Progress)
(5) Teays Valley SR-34 Corridor (Summer 2012)
(3) Elkins US-219 Corridor (In Progress)
(5) St. Albans US-60 Corridor (In Progress)
(8) Morgantown Beechurst/University Corridor (Spring 2013)
Rahall Transportation Institute is managing and evaluating
the performance of these systems
Adaptive Signal Control Evaluation

Morgantown WV-705 Corridor


Adaptive System: ACS-Lite & Traffic Responsive Hybrid
Evaluation Mechanism




Intersection Delay & Progression – Simulation using VISSIM
Travel Time – Bluetooth & GPS
Safety – Eventually crash records
Teays Valley SR-34 Corridor


Adaptive System: InSync
Evaluation Mechanism

Travel Time – GPS
WV 705 Signalized Intersections
~5.2 miles end-to-end
Stopbar detection
Advanced detection
16 Mainline Signals
WV 705 Traffic Generators
New Housing
School
Commuters
Residential
Shopping
Mall
Hospitals
WVU
Basketball
Research
Park
WVU
Football
Engineering
Campus
Shopping
Mall
Downtown
Campus
Student
Housing
CBD
School
WV 705 Corridor
WV 705 System Evaluation with Simulation

WV 705 corridor was modeled in VISSIM with traffic
signal controller simulators to evaluate:






Free operation (no coordination)
Basic Coordination Plans (TOD)
TOD + Traffic Responsive (TR)
TOD + ACS-Lite (ACS)
TOD + TR + ACS (Adaptive)
Total Delay and Hourly Delay Summarized



By System (All intersection movements)
By Mainline Movement (705 thru movements)
By Minor Movement (Left-turns and side-streets)
3000
2556
2526
All Movements
Total Delay (veh-h/h)
2500
2000
500
System Delay is lowest
in Free, but major
movements suffer
2236
9
9
7
6
7
6
1880
9
9
7
6
7
6
4
3
2
4
3
2
4
3
2
4
3
2
TOD
TOD+ACS
TR
TR+ACS
808
813
820
812
6
9
7
6
9
7
6
9
7
6
9
7
6
4
4
4
4
4
3
2
3
2
3
2
3
2
3
2
Free
TOD
TOD+ACS
TR
TR+ACS
1748
1714
1442
1424
9
7
6
9
7
6
4
3
2
4
3
2
TR
TR+ACS
1500
1000
2262
9
7
6
4
3
2
0
Free
1200
1021
Major Movements
Total Delay (veh-h/h)
1000
800
9
600
7
400
200
0
2000
Minor Movements
Total Delay (veh-h/h)
1800
1600
1400
1200
1000
9
9
7
6
7
6
9
7
6
4
3
2
4
3
2
4
3
2
Free
TOD
TOD+ACS
859
800
600
400
200
0
Minor Movements
4:30
5:00
5:30
6:00
6:30
7:00
7:30
8:00
8:30
9:00
9:30
10:00
10:30
11:00
11:30
12:00
12:30
13:00
13:30
14:00
14:30
15:00
15:30
16:00
16:30
17:00
17:30
18:00
18:30
4:30
5:00
5:30
6:00
6:30
7:00
7:30
8:00
8:30
9:00
9:30
10:00
10:30
11:00
11:30
12:00
12:30
13:00
13:30
14:00
14:30
15:00
15:30
16:00
16:30
17:00
17:30
18:00
18:30
Total Delay (veh-h)
Adaptive provides
most benefit over
TOD during off-peak
4:30
5:00
5:30
6:00
6:30
7:00
7:30
8:00
8:30
9:00
9:30
10:00
10:30
11:00
11:30
12:00
12:30
13:00
13:30
14:00
14:30
15:00
15:30
16:00
16:30
17:00
17:30
18:00
18:30
Major
40
Movements
35 suffer
in “free”30 mode
Total Delay (veh-h)
Major Movements
Total Delay (veh-h)
All Movements
100
90
80
70
60
50
40
30
20
10
0
FREE
TOD
TOD+ACS
TR
TR+ACS
Time of Day
25
20
FREE
15
TOD
10
TOD+ACS
5
TR
0
TR+ACS
Time of Day
60
50
40
30
FREE
20
TOD
10
TOD+ACS
TR
0
TR+ACS
Time of Day
Cumulative Total Delay –
University & Patteson
Cumulative Total Delay (veh-h)
Φ1: Westbound Left
Φ2: Eastbound
Φ3: Northbound Left
Φ4: Southbound
40
90
40
40
32
72
32
32
24
54
24
24
16
36
16
16
8
18
8
8
0
0
5:30
10:00
14:00
19:00
0
5:30
Φ5: Eastbound Left
10:00
14:00
19:00
0
5:30
Φ6: Westbound
10:00
14:00
19:00
5:30
Φ7: Southbound Left
90
40
40
32
72
32
32
24
54
24
24
16
36
16
16
8
18
8
8
0
5:30
10:00
14:00
19:00
0
5:30
10:00
14:00
19:00
19:00
0
5:30
10:00
14:00
Time of Day
Adaptive Off
14:00
Φ8: Northbound
40
0
10:00
19:00
5:30
10:00
14:00
Left-Turn Delay
Decreased by Shifting
Available Green Time
Adaptive On
28
19:00
Cycle Time
Coordination Diagram to Evaluate Offsets
GOOD
Vehicles arrive
after green
End of Phase
Vehicle Arrivals
BAD
Vehicles arrive
before green
Start of Phase
Time-of-Day
Offset Adjustment
Willowdale – Suncrest Town Center
ACS-Lite OFF
ACS-Lite ON
9000
8000
Cumulative Arrivals on Green
7000
6000
5000
4000
3000
2000
1000
0
0
3600
7200
10800
14400
18000
21600
25200
28800
Simulation Time
32400
36000
39600
43200
46800
50400
54000
Teays Valley WV-34 System Overview
5 Mainline Signals
Northbound - GPS Travel Time
Weekday 7AM–11AM
Distance Along Corridor (miles)
Average
(seconds)
35 mph Baseline
66
Before
175
After
110
% Change
-37%
Travel Time (seconds)
Before
After
Std Dev
(seconds)
-39
29
--
Northbound - GPS Travel Time
Weekday 11AM–2PM
Distance Along Corridor (miles)
Average
(seconds)
35 mph Baseline
66
Before
165
After
98
% Change
-41%
Travel Time (seconds)
Before
After
Std Dev
(seconds)
-33
29
--
Northbound - GPS Travel Time
Weekday 2PM–6PM
Distance Along Corridor (miles)
Average
(seconds)
35 mph Baseline
66
Before
168
After
100
% Change
-40%
Travel Time (seconds)
Before
After
Std Dev
(seconds)
-28
42
--
Planning for Traffic Signals




Most of the common planning software packages do not
account for basic traffic signal operations, much less
adaptive control
Difficult to evaluate/predict the performance of an
adaptive system to quantify the benefits
If there are fluctuating traffic patterns, adaptive might be a
good solution
If a signal system is being upgraded, often the incremental
cost for adaptive is insignificant
Contact Information
Andrew P. Nichols, PhD, PE
Director of ITS
Rahall Transportation Institute
Marshall University
Huntington, WV
304-696-3203
[email protected]