Adaptive Signal Control Technology: Bringing a Systems Approach to Implementation and Evaluation
October 18, 2011
Kevin Fehon, P.E., PTOE
DKS Associates
2
Overview
The past 40 years of adaptive signal systems The last five years in the USA Breaking out in 2011 Bringing a structure to system selection Performance Evaluation
First a definition
ASCT – Adaptive Signal Control Technology Terminology adopted by FHWA Used to refer generically to adaptive traffic signal
systems Avoids confusion with ACS-Lite
4
The past 40 years
It all started with SCATS in 1973 and SCOOT in 1975
Two fundamentally different but practically-oriented approaches
Both widely accepted outside the USA by 1985 Different approach attempted in USA, reflecting
different administrative and economic environment
5
SCOOT 2009
6
SCATS 2009
53 Cities; 179 Signal Regions
7
What has held up acceptance in the USA?
Agency capability and willingness System performance and benefits Lack of understanding Cost Availability of US-based systems
8
Growing acceptance in the USA.
ASCT is an important element of FHWA’s Every Day Counts (EDC) program
Increasing number of jurisdictions considering ASCT
Modern equipment with more processing power New ASCT systems emerging Still less than 1% of USA signals
9
US Adaptive Systems, April 2011
10
Systems currently available in USA
SCOOT/TACTICS SCATS InSync LADOT ATSC ACS-Lite OPAC KHAdence
QuicTrac RHODES Balance Synchro Green RHODES ATMS.now Delcan & Intellight
11
Introducing a structured approach to ASCT system selection
12
What is an appropriate definition of “adaptive”? Many different definitions are acceptable The definition you choose should be the one that
suits your circumstances How do you define real-time? Is it predictive or
reactive or both? What parameters are “adaptive”?
What are you trying to achieve with ASCT?
14
Different coordination strategies
Most coordination strategies can be categorized in one of the following four ways: Pipeline along a route (arterial or within grid) Progression in two directions along a route Equitable service for adjacent land uses Queue management
15
Non-coordination strategies
Some adaptive strategies do not involve coordination Optimal operation of an isolated intersection Optimal operation of a single critical intersection
within a coordinated area
Why use a systematic process?
Why SE?
Understanding the problem Managing risk
Projects getting bogged down with shifting requirements
Acquisitions being challenged by unsuccessful bidders/proposers/vendors
Projects not meeting agency needs (and it is mandatory for federal-aid projects)
18
Are you fully expressing your desires?
Is ASCT really what you need? Can you clearly document the deficiencies of
your existing system that you wish to overcome with adaptive?
Can you clearly document the existing capabilities of your system that must be retained with adaptive operation?
19
Can you clearly verify that you got what you asked for? Are your needs clearly expressed? Is each need translated into a requirement Is every requirement actually satisfying an
expressed need? How is each requirement tested, so the that the
vendor can verify it is satisified?
20
Can you validate that you got what you needed? Have your organization’s goals and objectives
been clearly defined? Are measures of performance clearly defined? Can you test the operation against the measures
of performance?
What is the purpose of the Model Systems Engineering Documents?
22
Provide a framework to guide agencies to the most appropriate solution Provide sample statements that can be used in a
Concept of Operation Provide sample requirements that can be used to
select/purchase/design a system Describe appropriate verification and validation
plans
How should your process work?
24
Overview
First pass in half to one day, by experienced agency staff
Identify additional information you need in order to make appropriate decisions and document the basis of those decisions
25
Work Flow
26
What are the pitfalls of not following a structured approach?
28
Pitfalls
You don’t get the most suitable system You may lose some capabilities You may not get the system you want Making a constraint a requirement at the outset
distorts the outcome. You can’t quantify the trade-offs you are making
29
The structured approach will result in the most appropriate ASCT selection Required capabilities will be clearly stated Assumptions will be explicit, so requirements
won’t fall through the cracks Trade-offs to satisfy real and perceived
constraints will be explicit and justified Expectations are more likely to be met
30
Is it worth the effort?
Emphatically yes! Some ASCT systems that have not fulfilled
expectations are falling by the wayside The established systems have strongly
documented benefits and high B/C ratios Much misinformation, both positive and negative
Gresham PM Peak TT(Eastbound Burnside)
405
368
424393
349 373
314
0
100
200
300
400
500
Trav
el T
ime
(sec
)
1997 Free
1998 TOD
2004 Free
2004 (1998 TOD)
2004 TOD
2007 (2004 TOD)
2007 SCATS
Legend
32
Evaluation techniques often hide the truth
Travel time per direction does not show system-wide effects, and confuses positive and negative impacts.
Travel time per period do not show effects of varying demand levels.
Volume-weighted, system-wide evaluation allows differentiation of conditions.
33
Sunnyvale: System Expansion
34
Questions?