U.S. Department of TransportationFederal Highway Administration

ACS-LiteFHWA Adaptive Signal

Control Systems

Raj S. Ghaman, P.E.Team Leader, Office of Operations RD & T

Traffic Signal & Operations Working Group March 26, 2008

Washington D.C., USA

U.S. Department of TransportationFederal Highway Administration

Slide 2

Outline

• Goals

• System architecture

• Adaptive approach– Cyclic performance measures

• Field trials

U.S. Department of TransportationFederal Highway Administration

Slide 3

Adaptive in the U.S. (FHWA)

• 1970s-1980s: UTCS– Second by second central

• 1990s: Predictive control (ACS)– FHWA ACS: RHODES, OPAC– Second by second distributed

• 2002: Controller-based Adaptive– FHWA ACS “Lite”: Siemens ITS– Leverage existing hardware– Update controller parameters every five minutes

U.S. Department of TransportationFederal Highway Administration

Slide 4

FHWA Goals for ACS-Lite

• Low cost design

• Leverage existing infrastructure• Standard US-style actuated controllers (rings,

phases, splits, barriers)• Standard fully-actuated detector layouts• Standard NTCIP Communications

• “Retro-fit” with major US signal system vendors

U.S. Department of TransportationFederal Highway Administration

Slide 5

Project Team

U.S. Department of TransportationFederal Highway Administration

Slide 6

System Architecture

Vendor

Field Master

Protocol Translation

9600bps, up to 12 controllers

NTCIP + ACS-Lite firmware upgrade

Vendor Specific or NTCIP

NTCIP

ACS-LiteOptional

Optional

Field Processor

U.S. Department of TransportationFederal Highway Administration

Slide 7

ACS-Lite Detection Layout

Flow ProfileDetectors

Phase Utilization Detectors

Need detectors at stop-bar of coordinated phases

U.S. Department of TransportationFederal Highway Administration

Slide 8

21

ACS-Lite adaptive control philosophy

• Data-driven parameter tuning

• Limited/no traffic modeling

• Recent past predicts the near future• Splits

– Phase Utilization

• Offsets– Statistical Flow Profiles

U.S. Department of TransportationFederal Highway Administration

Slide 9

ACS-Lite NTCIP firmware upgrade

• Phase Timing Status Object

• Detector Status Object

• Configuration Objects• Polled once per minute

– Second-by-second accuracy– Bandwidth efficient

• Minute-by-minute polls are “stitched” together for cycle-by-cycle performance assessment

1

2

3

U.S. Department of TransportationFederal Highway Administration

Slide 10

Occupancy per phase interval

• Occupancy values per second

• Correlated to Red/Green/Yellow

37s 4s49s

Second-by-secondOccupancy

37s 4s32s

Second-by-secondOccupancy

Cycle 1

Cycle 2

17s

U.S. Department of TransportationFederal Highway Administration

Slide 11

Cycle-by-cycle Data

Phase Timing Volume/Occupancy

U.S. Department of TransportationFederal Highway Administration

Slide 12

Occupancy per interval split tuning

57s4s29s

Phase 2AverageOccupancy

Unoccupied

Averaging

U.S. Department of TransportationFederal Highway Administration

Slide 13

Green occupancy Phase Utilization

U.S. Department of TransportationFederal Highway Administration

Slide 14

Cyclic occupancy profiles Statistical profile

“statistical” flow profile

Example shows need to move offset so green

corresponds with traffic earlier in cycle

U.S. Department of TransportationFederal Highway Administration

Slide 15

Progression Performance Offset tuning

• Performance measure for offsets “capture efficiency”

• Shift offsets small amount

• Constrain changes within user-configurable bounds

Inbound

Outbound

U.S. Department of TransportationFederal Highway Administration

Slide 16

Field trials

PEEKSt. Petersburg/Tampa, Florida

EconoliteColumbus, Ohio

Eagle/SiemensHouston, Texas

McCainSan Diego, California

U.S. Department of TransportationFederal Highway Administration

Slide 17

Field trial – Columbus, Ohio

• Early “lessons learned”– Communications integrity– Detector configuration

• Separate channels per lane

– Remote configuration capability

– Details, details

High School

Freeway

3km

Major road

Major road

Construction zone

U.S. Department of TransportationFederal Highway Administration

Slide 18

Columbus, Ohio - Hamilton Road

Study Area N

U.S. Department of TransportationFederal Highway Administration

Slide 19

Columbus, Ohio – Hamilton Road (con’t)

U.S. Department of TransportationFederal Highway Administration

Slide 20

ACS-Lite Field trial – Houston, Texas State Route 6

Study AreaN

U.S. Department of TransportationFederal Highway Administration

Slide 21

Houston, Texas – State Route 6 (con’t)

U.S. Department of TransportationFederal Highway Administration

Slide 22

ACS-Lite Field trial – Tampa, Florida State Route 70

Study Area

N

U.S. Department of TransportationFederal Highway Administration

Slide 23

Tampa, Florida – State Route 70 (con’t)

U.S. Department of TransportationFederal Highway Administration

Slide 24

ACS-Lite Field trial – San Diego, California Main Street

N

Study Area

U.S. Department of TransportationFederal Highway Administration

Slide 25

San Diego, California – Main Street (con’t)

U.S. Department of TransportationFederal Highway Administration

Slide 26

Future

• Algorithms enhancements (2008):– Long-term parameter

adjustment • Seasonal baseline parameters• TOD schedule switch points

– Cycle time tuning– Selection of transition

method– Better Graphical User

Interface (GUI)

U.S. Department of TransportationFederal Highway Administration

Slide 27

Questions?