1 of 35

1

Presented by

Sanjay Paul, PhD, PE, PTOE, PTP, RSP

Desert Southwest Area Traffic Manager, HDR

Technologies and Tricks:

Taking HAWKs to the Next Stage

Simon Ramos, PE

Field Services Superintendent, City of Phoenix

ITS Arizona Annual Conference

October 20, 2022

2 of 35

2

46% Pedestrian- Vehicle Crash Reduction

Phoenix has 81 HAWKs

3 of 35

3

Safety Issues

4 of 35

4

Elderly Population

COVID Effect

Large Party Size

5 of 35

5

Side Street Delay- �Coordinated Signals

6 of 35

6

Many DOTs Set Up Their Detection for Fully Actuated Control But Set Up Their Controller for Semi-Actuated Control

Coordinated Phases Always Force-Off in Semi-Actuated Control

Reduce Side-Street Delay

Coord Actuated/Early Coord Gap-Out

7 of 35

7

Reduce Side-Street Delay-Yield Point

Semi-Actuated Control

Yield Point - A point in a coordinated signal operation that defines where the controller decides to terminate the coordinated phase.

8 of 35

8

Reduce Side-Street Delay

Fully-Actuated Control

If the coordinated phases are actuated, they will extend. The amount of extension is controlled by the coordinated phase split extension interval. If the coordinated phase are no longer actuated, they will gap out provided there is demand in any of the following phases in the ring sequence.

9 of 35

9

Reduce Side-Street Delay

= Side Street Split

​

=Main Street Split

Reversed Split Timing – Typically 2-Phase

Example: 120s CL

10 of 35

10

Reversed Split Timing – Typically 2-Phase

~40s Band

Reduce Side-Street Delay

11 of 35

11

Reduce Side-Street Delay

Reversed Split Timing – Typically 2-Phase

12 of 35

12

Reduce Side-Street Delay

Reversed Split Timing – Typically 2-Phase

13 of 35

13

Reduce Side-Street Delay

Reversed Split Timing – Typically 2-Phase

Side Street Calls

(Random Events)

14 of 35

14

Reduce Side-Street Delay

Reversed Split Timing – Typically 2-Phase

Side Street Calls

(Random Events)

15 of 35

15

Reduce Side-Street Delay

= Side Street Split

​

=Main Street Split

16 of 35

16

  • Increasing HAWK Efficiency
    • Reducing Pedestrian Delay
    • Reducing Pedestrian Violations
    • Increasing Pedestrian-Vehicular Safety

  • Focus users: Pedestrians/Bicyclists

(+ vehicles)

  • Pilot deployment at 6 locations
    • New Signal Timing Scheme
    • Advanced Vehicle Detection
    • Detection & Green Extension
    • Passive Pedestrian Detection

Study Objectives & Scope

17 of 35

17

Fillmore/7th Ave

Thomas/2nd St

Montebello/7th St

32nd St/Grand Canal

McDowell/Grand Canal

24th St/Grand Canal

Advanced detections were deployed along Grand Canal

​

Study HAWK Locations

New timings implemented at:

      • Downtown
      • Midtown
      • Uptown

18 of 35

18

Camera & Visions

19 of 35

19

Technology Testing and Data Collection

Before & After Sunrise

Before & After Sunset

Mid-day

20 of 35

20

Study Methodology –

Passive Detection Evaluation

    • Passive Detection and Pedestrian Call Capability
      • Without pressing the button, video camera detects and puts pedestrian call into the controller
      • 5-6 seconds in the detection zone

21 of 35

21

Study Methodology –

Dynamic Detection Evaluation

    • Detection and Clearance Extension Capability
      • Ped interval + clearance
      • MUTCD/common practice – 3 ft/sec to 4 ft/s
      • ITE studies – 1.8 ft/sec to 8 ft/sec
      • Call to hold conflicting phase for 7 seconds

22 of 35

22

Delay Estimation and Effectiveness Evaluation

  • n1 - sample size from sample set 1
  • s1 - sample standard deviation of sample set 1
  • n2 - sample size from sample set 2
  • s2 - sample standard deviation of sample set 2

Inference for Independent Means

Independent two sample set

  • df – degrees of freedom
  • t = t- statistics

23 of 35

23

Impact of New Signal Timing Scheme

on Pedestrian Delay

50s

24s

Before

After

11-ish

Average Pedestrian Delay

24 of 35

24

Impact of New Signal Timing Scheme

on Pedestrian Delay

48%

75%

54%

32%

52% Reduction

25th

50th

Avg

75th

100th

25 of 35

25

Impact of Advanced Detection

on Pedestrian Delay

41s

21s

Before

After

11-ish

Average Pedestrian Delay

26 of 35

26

Impact of Advanced Detection

on Pedestrian Delay

9%

39%

76%

1%

48% Reduction

25th

50th

Avg

75th

100th

27 of 35

27

Comparisons of Vehicle Delay

New Logic & Advanced Detection

41s

47s

43s

43s

Before

After

Before

After

New Timing

New Timing + Advanced Detection

Average Vehicular Delay

28 of 35

28

Passive Detection Accuracy: 89%

Summary: Passive Detection

29 of 35

29

Dynamic Detection & Extension Accuracy: 80%

Summary: Dynamic Detection & Crossing Time Extension

30 of 35

30

  • Both new timing scheme and advanced detections are effective in reducing pedestrian delay without increasing vehicular delay
  • This method can be adopted for signal operations (will have some drawbacks to overcome)
  • Tremendous improvements in algorithms and technologies in recent years
  • Two cameras to cover both directions, separate camera for passive

detection, upgraded controller is needed to have all 4 ports working

  • Proper intersection lighting & wider detection zone are critical
  • Passive detection needs appropriate ramp marking and signage in full implementations
  • Available advanced pedestrian and bicyclist features:
      • Walking or biking speed
      • Pedestrian delay at crossings
      • Pedestrian violations

Summary of Findings

31 of 35

31

Challenges in Pilot Deployment and Testing

  • Coordination with multiple entities
      • City of Phoenix TMC
      • City of Phoenix Signal Shop
      • City of Phoenix IT department
      • ITERIS/Technology Vendor
  • Camera angle and coverage on main line and cross-walks Supplementary Camera was added for better coverage
  • Communication gap between sales personnel and software development engineers
  • Communication gap among software development engineers/data scientists
  • Controller could not handle four ports/over night shipping
  • Access & Communication issues
  • Arizona/Phoenix heat, 109o F in September

32 of 35

32

Sanjay Paul

Subhankar Dey

Miranda Sundblom

Mike Barton

Special thanks to

Bruce Littleton

Deputy Director, Street Transportation

City of Phoenix

Maddison Klingberg

ITERIS Technology

Simon Ramos

Eric Hernandez

Mailen Pankiewicz

Robert Lugo

Thanks to the

Project Team

33 of 35

33

Thank you very much!!

Comments and questions

are welcome

Contacts:

​

Simon Ramos, PE

Simon.Ramos@phoenix.gov, 520.500.4190

​

Sanjay Paul, PhD, PE, PTOE, PTP, RSP

Sanjay.Paul@hdrinc.com, 480.559.2623

​

34 of 35

34

Impact of New Signal Timing Scheme on Pedestrian Delay (cont’d)

% - Before - After

25% - 21s - 11s

50% - 44s - 11s

75% - 1:14 - 34s

100% - 2:06 - 1:26

35 of 35

35

Impact of Advanced Detection

on Pedestrian Delay (cont’d)

% - Before - After

25% - 11s - 10s

50% - 18s - 11s

75% - 1:07 - 16s

100% - 2:08 - 2:07