Title: Tools and Methodology for CorridorLevel Traffic Operations Planning
1Tools and Methodology for Corridor-Level Traffic
Operations Planning
- California Center for Innovative Transportation,
November 2007
2About CCIT
- The California Center for Innovative
Transportation (CCIT) accelerates the
implementation of research results and the
deployment of technical solutions by
practitioners to enable a safer, cleaner and more
efficient surface transportation system.
3CCITs Vision
CCIT envisions a world where
4Acknowledgments
- Hamed Benouar (CCIT)
- R. Jayakrishnan (UC Irvine)
- Steve Hague (Caltrans)
- Fred Dial (Caltrans)
- Henry Liu (UMN)
- Tarkek Hatata (SMG)
- Jeff Ban (CCIT)
- Erik Alm (CCIT)
- Lianyu Chu (CCIT)
- Jeff Ban (CCIT)
System Metrics Group Inc.
5A picture speaks a thousand words
Source Performance Measurement System (PeMS)
October 2001 Vphpl volume per lane per hour
6Outline
- The CCIT Corridor Management Project
- The Case for Micro-Simulation
- Selected Results and Outcomes
- Additional Takeaways
7The Corridor Management Modeling Project
- 3-year project 2004-2007
- Major Tool Microscopic Traffic Simulation
(Paramics) - Outcome
- 3 Corridor Studies Lessons Learned
- I-880 near San Francisco
- SR-41 in Fresno
- I-5 in Orange County
- A template for corridor traffic operations
planning
8I-5 Corridor
30 miles of I-5 and nearby arterials 200
signals 70 ramp meters
9I-880 Corridor
30 miles of I-880, parallel arterials 190
signals 69 ramp meters
10SR-41 Corridor
16 miles of SR-41, 2 parallel arterials 90
signalized intersection 15 ramp metering
controllers
11Large Network/Corridor Applications
- Evaluate traffic management strategies and
operational improvements - Calibrate / optimize operational parameters of
ITS strategies - Develop / test new models, algorithms, control
strategies - Can even be used for project selection, including
capital projects
12Overall Approach
13Corridor Studies as part of a Global Systems
Management Framework
14Outline
- The CCIT Corridor Management Project
- The Case for Micro-Simulation
- Selected Results and Outcomes
- Additional Takeaways
15Why Micro Simulation?
- Analysis of traffic flows
- Reasonable representation of queues and resulting
traffic congestion - Focus on bottlenecks
- Ability to analyze operational projects
- Integration of planning and operations
- Micro-simulation-based presentations/visuals
effective tools for discussions with stakeholders
16MicroSim vs. Travel Demand Model
17Micro Simulation Validation Criteria
18Calibration Validation Issues
- The OD estimation step is most important in model
validation - The time resolution used in the OD estimation is
crucial - For instance, if observed data and speed plot
have one-minute resolution, estimating OD tables
at 5 min resolution may be OK, but certainly not
more than 15 min - Any OD estimation scheme that uses multiple time
period static assignments within the iterative
scheme is troublesome - Also, some of the ad-hoc and supposedly pragmatic
schemes such as adjusting the vehicle flows
during the simulation to match the observed
counts are very questionable - The effect of the initial seed matrix could be
significant in the OD estimation process - Matching counts with hourly GEH values is
questionable - Maybe insufficient if the intent is to model with
5-minute accuracy - No reasonable current criteria on how speed-plots
are to be matched - What traffic condition (representative day) to
calibrate against? - Calibrating against averaged conditions is
dangerous and may be impossible - Travel time calibration Usually floating car
data is expensive, and the variation of observed
travel time on a section-basis is very high
19Calibration Observed day-to-day Variation
An analysis of two years of data on I-880 showed
ONE day without any accidents reported by the
California Highway Patrol. What is the
representative conditions to calibrate against?
20Calibration Average Data?
Averaged observed conditions may lead to
infeasible conditions (i.e., No OD matrix can
cause such traffic Forward growing queue etc!)
21Outline
- The CCIT Corridor Management Project
- The Case for Micro-Simulation
- Selected Results and Outcomes
- Additional Takeaways
22Speed contours from simulation
23SR-41Scenarios
- Add auxiliary lanes
- On-ramp widening
- Ramp metering systems
- Left-turn lanes at intersections
- Add new traffic signals
- Expand freeway from 6 to 8 lanes
- HOV lane (optional)
- The Short-term Improvements Alternative (year
2005) - The Medium-term Improvements Alternative (year
2010) - The Long-term Improvements Alternative (year
2025), and - The Long-term Improvement Alternative with HOV
lanes (year 2025)
24SR-41 Scenarios (2)
25Corridor Systems Management Plan Template
26Outline
- The CCIT Corridor Management Project
- The Case for Micro-Simulation
- Selected Results and Outcomes
- Additional Takeaways
27I-880 Average Daily Delays
28Congestion by Cause
29Bottleneck Analysis
NORTHBOUND I-880 AVERAGE SPEED (mph) 10/1/04
10/31/04 (TUE-THUR)
Recurrent traffic congestion
bottleneck location
23rd Avenue
SR-238
length of congestion
Tennyson
DIRECTION OF TRAVEL
Fremont
period of congestion
Auto Mall Pkwy
5 AM
11 AM
5 PM
11 PM
12 AM
TIME OF DAY
SPEED COLORS (mph)
25
30
35
40
45
50
55
60
65
70
75
30Bottleneck Variations
31Bottleneck Variations (B)
32Bottleneck Variations (C)
33Conclusions
- Micro-Simulation is EXPENSIVE and RISKY
- Considerable work involved
- Calibration and validation issues
- Priorities reversed modeling takes over
objectives - Remains the holy grail for operations planning
- A corridor-wide, system-level approach goes a
long way - Systematic data collection
- Root-cause analysis
- Project prioritization
- Tools are available
- Performance Measurement System (PeMS)
- CSMP Template associated Cookbook
34Tools and Process for Corridor-Level Traffic
Operations Planning
QA
- California Center for Innovative Transportation
- November 2007