View
245
Download
0
Embed Size (px)
DESCRIPTION
Nebraska Department of Roads Conference 2012 Presented by Dr. Larry Rilett
Citation preview
Evaluation of NDOR’s Active Advance Warning System
Laurence R. Rilett Ph.D., P.E.University of Nebraska-Lincoln
Presentation Outline
• Background• Analyses
– Safety– Operation– Simulation– Sensitivity
• Conclusions• Recommendations
Background• Dilemma Zone
– At the legal speed limit, the driver can neither clear the intersection before the end of the intergreen period nor stop without entering the intersection.
Background• Dilemma Zone: NDOR 2002 Report
– “Length of roadway in advance of the intersection wherein drivers may be indecisive or respond differently to the onset of the yellow indication.”
– Also known as “option zone” or “zone of indecision”
Background
• If an intersection is designed correctly (e.g. NDOR) a dilemma zone will not exist– Assuming deterministic system
• Vehicles same characteristics (accelerate, decelerate, weather, etc.)– Trucks/braking
• Drivers make the correct decisions– Stop, proceed
• Assuming: legal maneuvers (not running red light)
Potential Problems
• A major safety concern at high speed signalized intersections
Common Treatments
• Advance Warning (AW) Flashers– Flashing signal heads and warning signs
• Activated at predetermined time before end of green
• “Mixed” results regarding effectiveness
Common Treatments• Advance Detection (AD)
– Series of detectors in advance of intersection• Extend green on detection
– Effective in reducing crashes and conflicts– Increases likelihood of extending green to maximum
(max-out)• Dilemma zone protection is lost
NDOR’s Actuated Advance Warning (AAW) System • Combines advance detection and advance warning
– Single detector– Shorter maximum allowable headway– Lower frequency of max-out
Issues
• Results positive but mostly anecdotal• Guidelines for installation
– When do they need to be removed (if ever)?• Motivation for study
Part 1
Crash Data Analyses
Safety Effectiveness
• Test Sites– 26 treated intersections– 29 reference intersections
• “Similar” characteristics as treated intersections
• Provided by NDOR– 13 year of crash counts and AADT
• 1996-2008
Treated Intersections: Table 2.2
Simple example ignores regression to mean, changes in AADT…Need to compare to untreated intersections…
Safety Effectiveness
• Method– Full Bayes– Accounts for uncertainty in data– Generates a distribution of likely expected number of
crashes– Combines this distribution with site-specific crash
data to obtain expected crash frequency– Approach is complex but requires less data
Safety Effectiveness
• Crash Reduction Rate
Safety Effectiveness
• Model
Safety Effectiveness Results
Safety Effectiveness Results
Safety Effectiveness Results
Safety Effectiveness Results
Safety Effectiveness Results
Safety Effectiveness Results
Safety Effectiveness Results
Safety Effectiveness Results
Part 2
Operational Analyses
Operational Analyses
• Main Characteristics– Approach speeds– Acceleration/deceleration characteristics
• Following onset of yellow• During lead flash
– Frequency of max-outs– Rate of dilemma zone “entrapment”– Waiting time on conflicting phases
Study Site: Lincoln
• Highway 77 and Saltillo Road
Study Site: Omaha
• Highway 370 and N 132nd Street
Operational Analyses
• Data
Operational Analyses
• Max-out probabilities
Operational Analyses
• Waiting time on minor road
Operational Analyses
• Waiting time on minor road
Lincoln (Figure 3.13)
• Acceleration/deceleration- lead flash
Omaha
• Acceleration/deceleration- lead flash
Operational Analyses
• (Average) speed profile- lead flash
Operational Analyses
• Acceleration/deceleration- yellow
Operational Analyses • Acceleration/deceleration- yellow
Operational Analyses
• (Average) speed profile- yellow
Operational Analyses
• Vehicles in dilemma zone- yellow
Part 3
Microsimulation Analyses
Microsimulation Model
• VISSIM– Inputs: geometry, traffic counts, timing, speeds, etc.
• Calibration– Adjust model parameters such that field data
“matches” simulated data– Measures of performance
• Average waiting time• Speed profile
Microsimulation Model • GA Calibration Procedure
Microsimulation Model
• Calibration Results
Microsimulation Model
• Measures of performance– Average waiting time
Microsimulation Model
• Measures of performance– Speed profile
Sensitivity Analysis • Experimental Design
Sensitivity Analysis
• Simulation runs– 480 total factor combinations– 1-hour simulation run for each– 10 replications each
• Output– Waiting times– Number of conflicts
Sensitivity Analysis
• Effect of turn percentage– On average waiting times
Conclusions
• Safety effects– Greater than 90% probability that installation of
system is beneficial• Operational effects
– Lower than expected number of vehicles in dilemma zone
– Low max-out probabilities– System seems to work well
Conclusions
• Simulation model– Developed framework for modeling system– Successfully applied to two sites
• Sensitivity analysis – Site specific– Can be used to perform sensitivity analyses
Recommendations
• System worth considering at other high-speed signalized intersections– From a safety perspective
• Guidelines regarding installation– McCoy and Pesti (2002)
• Guidelines regarding removal– Simulation study
• Max out, delay, etc.
Slide design © 2009, Mid-America Transportation Center. All rights reserved.
Dr. Laurence Rilett, Ph.D., P.E. University of Nebraska-Lincoln
CREDITS