Britton Hammit, Kimley-HornJiaxin Tong, Kimley-HornSanhita Lahiri, VDOT
VDOT VISSIM User Guide
2020 SimCap Meeting
January 14, 2020
Outline
• Introduction
• Overview of Contents
• Vissim 11 Evaluation
• Calibration Guidance
• Conclusions
2
Purpose and Need• Supplement VDOT Traffic Operations and Safety
Analysis Manual (TOSAM).
• Provide guidance and Virginia best practices for model development and analysis.
• Maintain consistency for peer and agency review.
• Promote the quality and consistent VISSIM analysis in support of decision making for project planning and infrastructure.
3
User Guide Contents• Model Development
• Model Review and Debugging
• QC Checklist
• Results and Preparation
• Calibration Guidance
• Model Scenarios
4
Vissim 11 Evaluation
5
Vissim 11 EvaluationObjectives• Improve understanding of Vissim 11 to
provide VDOT recommendation on its adoption for new projects.
• Provide guidelines for running legacy models in Vissim 11.
• Highlight differences in parameter sensitivities between Vissim 11 and legacy software versions.
• Update existing Vissim calibration guidance.6
Questions of Interest1. Will Vissim models calibrated in legacy
Vissim versions remain calibrated when run directly in Vissim 11?
2. Do these driving behavior model changes elicit a change in calibration guidance for specific parameters?
7
Legacy Vissim Models• Compare the results of previously calibrated
models in prior versions of Vissim with the results obtained by running those models in Vissim 11.
• 10 Calibrated Models (5 Networks, AM/PM)• Models selected with varying network sizes,
network features, and operational conditions.
8
Legacy Vissim Models
9
Name Maintained Calibration in Vissim 11
Primary MOE Difference between
Versions
Severity of MOE Difference
Route 7 AM No Queue Length Moderate
Route 7 PM No Travel Times Substantial
Route 28 AM Yes - -
Route 28 PM No Freeway speeds Substantial
Seminary Road AM No Freeway speeds Substantial
Seminary Road PM Yes Freeway speeds Moderate
I-95 AM Yes - -
I-95 PM No Speed Criteria Moderate
Route 123 AM No Travel Times Substantial
Route 123 PM No Travel Times Substantial
Potential Causes• Unique traffic conditions
• Model development
• Model calibration
10
Model Development
AM Model Calibration
PM Model Calibration
Calibration
11
CalibrationAdjust these:
12
CalibrationUsing these:
13
Freeway Car Following Model (Wiedemann 99) – Calibration Parameters
Parameter Default Value Unit
Suggested Range
Basic Segment
Weave/Merge/Diverge Segment
CC0 Standstill distance 4.92 feet (ft) 4.5 to 5.5 >4.92
CC1 Headway time 0.9 seconds (s) 0.85 to 1.05 0.90 to 1.50
CC2 ‘Following’ variation 13.12 ft 6.56 to 22.97 13.12 to 39.37
CC3 Threshold for entering ‘following’ -8 -- Use default
CC4 Negative ‘following’ threshold -0.35 -- Use default
CC5 Positive ‘following’ threshold 0.35 -- Use default
CC6 Speed dependency of oscillation 11.44 -- Use default
CC7 Oscillation acceleration 0.82 ft/s2 Use default
CC8 Standstill acceleration 11.48 ft/s2 Use default
CC9 Acceleration at 50 mph 4.92 ft/s2 Use default
Arterial Car Following Model (Wiedemann 74) – Calibration Parameters
Parameter Default Value Unit Suggested
Range
Average standstill distance 6.56 feet (ft) 3.28 to 6.56
Additive part of safety distance 2.00 -- 2.0 to 2.2
Multiplicative part of safety distance 3.00 -- 2.8 to 3.3
CalibrationTo reach these:
14
Calibration Styles
15
Calibration Styles
Differences in Calibration Style
• Activation of the Cooperative Lane Change parameter at merge, diverge, and weaving segments.
• A low Safety Distance Reduction Factor.
• Activation of the Advanced Merge parameter.
16
Cooperative Lane Change
17
Cooperative Lane Change
Activated
Calibrated
Route 28 AM
I-95 AM
Not Activated
Uncalibrated
Route 28 PM
I-95 PM
Cooperative Lane Change
18
Route 28, AM ModelCLC Activated - Calibrated
Field Vissim 9 Vissim 11 Field Vissim 9 Vissim 11
Route 28, PM ModelCLC not Activated – Not Calibrated
Safety Distance Reduction Factor
19
Safety Distance Reduction Factor (SDRF)
0.25
Uncalibrated
Seminary Road AM
Route 123 AM
Route 123 PM
0.55
Calibrated
Seminary Road PM
Safety Distance Reduction Factor
20
Seminary Road, PM ModelSDRF Near Default – Calibrated
Field Vissim 9 Vissim 11
Field Vissim 9 Vissim 11
Seminary Road, AM ModelSDRF Significantly Below Default
– Not Calibrated
Advanced Merge
21
Advanced Merge
Activated
Calibrated
Route 28 AM
Seminary Road PM
I-95 AM
Not Activated
Uncalibrated
Route 28 PM
Seminary Road AM
I-95 PM
Summary• Models that were previously calibrated in legacy
versions of Vissim were more likely to maintain calibration in Vissim 11 when
• Cooperative Lane Change and Advanced Merge features were activated.
• Safety Distance Reduction factor was maintained at a conservative value near default.
• When converting a calibrated legacy model to Vissim 11, time and resources should be allotted to re-visit model calibration and fully QC model operations.
22
Questions of Interest1. Will Vissim models calibrated in legacy
Vissim versions remain calibrated when run directly in Vissim 11?
2. Do these driving behavior model changes elicit a change in calibration guidance for specific parameters?
23
Calibration Parameter Evaluation• Sensitivity analysis of parameters in Vissim 9
and 11 to study the impact of changes in parameter values on model outputs
24
5 Models
17 Parameters
4 Parameter Values
Summary of Findings• Updates to lane change algorithms in Vissim 11
should yield smoother traffic flow under the default parameter settings.
• Thus, modelers should need to make fewer changes towards a more “aggressive” driving behavior settings.
• Safety Distance Reduction Factor• Lane Change Distance
• No evidence that previously-established recommendations for parameter ranges need to be revised.
25
Findings Practical GuidanceEvidence that some “Calibration Styles” arepreferred over others.
• Narrowing ambiguity in the calibration process.• Improving resulting model robustness,
transferability, and consistency.
26
Calibration Guidance
• Calibration parameter definitions
• Calibration plan
• Pre-Calibration set-up
• Calibration guidelines
• Calibration steps and structure27
Calibration Guidelines1. Activate cooperative lane change at all merge,
diverge, weaving segments.
2. Keep freeway car-following model parameters CC0 and CC1 consistent between all connected freeway links.
3. Freeway car-following model parameter CC2should be greater in merge, diverge, weaving segments compared to basic segments.
4. Safety Distance Reduction Factor adjustment should be a last resort in model calibration.
28
CC1 Consistency
29
CC1 = 0.9 sec
CC1 = 1.1 sec
CC1 = 1.2 sec
30
31
Round 1Iterative adjustments of:• CCO/CC1• CC2 for “basic segments”• CC2 for “merge, diverge, and weaving segments”• Lane Change Distance
Round 2Iterative adjustments of:• Additional seeding conditions• Lane change deceleration parameters• Cooperative Lane Change parameters• Minimum Net Headway lane change parameter
Round 3Iterative adjustments of:• Addition of special Driving Behavior Containers• Safety Distance Reduction Factor lane change parameter
Freeway Corridor Calibration
Only if necessary
Summary
32
• Model calibration will always be an iterative, trial-and-error procedure with numerous uncertainties.
• VDOT’s user guide introduces new guidelines and steps that can be used to structure this iterative procedure to improve consistency, efficiency, and model validity.
VISSIM User Guide is available at http://www.virginiadot.org/business/resources/VDOT_Vissim_UserGuide_Version2.0_Final_2020-01-10.pdf
Questions?
Britton Hammit, [email protected]
Jiaxin Tong, [email protected]
Sanhita Lahiri, [email protected]
33