Lexington Area TransCAD Travel Demand Model

1
Lexington Area TransCADTravel Demand Model

• July 22, 2003
• Kentucky Traffic Model Users Group Meeting
• Kyeil Kim, Ph.D.
• Bernardin, Lochmueller Associates, Inc.

2
Main Features

• Consists of separate time-of-day (TOD) models
• Realistic free-flow speed based on a speed survey
  and HCM-based signal delays
• Varying capacities by TOD for reversible lanes
• Model parameters are based on recent travel
  survey and KYTCs HIS database.
• 2-stage assignments Initial and Feedback
• User-friendly model GUI

3
Model Area
4
Roadway Network

• Incorporation of KYTCs HIS and
  geometric/operations data
• All MINUTP network attributes
• FHWA functional class, posted speeds, etc.
• Number of lanes, reversible lanes bike lanes
• Area type, shoulder widths, median type, etc.
• Traffic signals multi-way stops (signal
  priority, multiple signals)
• Turn prohibitors

5
TAZ DB

• Total 489 zones
• Internal zones 445 zones
• External zones 44 zones
• 51 data attributes
• Date related to population and household
• Employment by SIC code
• Student population, group quarters school
  enrollment
• Vehicle ownership
• Key data for estimating trip productions and
  attractions

6
Speed/Capacity Estimation

• Free-flow speed, peak-hour capacity daily
  capacity
• A special GIS-dk program to estimate directional
  capacities by TOD free-flow speed
• Consideration of changes in reversible lanes by
  TOD
• Free-flow speed f (functional class, posted
  speed, number of lanes,
  facility type, delays on
  interrupted facilities)
• Daily capacity f (functional class, number of
  lanes, facility type)
• Peak-hour capacity max. service flow
  adjustment factors
• where, Adjustment factors f (geometric data,
  functional class, facility
  type, area type, delays on interrupted
  facilities)

7
Speed/Capacity Estimation (Contd)

• Delays on interrupted facilities
• HCM 2000 procedure
• delay/veh uniform delay PF incremental
  delay initial queue delay
• where, PF progression factor f
  (arrival type, g/C)
• uniform delay
• Varying g/Cs and PFs by signal priority and
  multiple signals

8
Lexington ModelGUI
9
Lexington ModelGUI
10
Trip Generation

• Six internal trip purposes
• - HBW, HBK12, HBU, HBO, NHBW NHBO
• Trip production
• Cross-classification technique
• Data The 2000 Knoxville Household Travel
  Behavior Study,
• The 1999 Indiana University Travel Demand
  Survey
• Analysis of Variance (ANOVA) and non-parametric
  correlations techniques to identify the
  predictor variables for various trip purposes
• Stratification curve distribution of
  households in a zone over various
• levels of the predictor variables

11
Trip Generation (Contd)
Trip production model
Trip Purpose 1st Predictor 2nd Predictor Overall Trip Rate
HBW Workers/H.H. Vehicles/H.H. 1.30
HBK12 Students/H.H. None 0.64
HBU University Student/TAZ None 0.48
HBO Household Size Vehicles/H.H. 3.72
NHBW Workers/H.H. Household Income 0.84
NHBO Household Size Vehicles/H.H. 2.02
Total Total Total 9.00
12
Trip Generation (Contd)
Trip attraction model the Atlanta regression
model
Trip Purpose Independent Variable Parameter
HBW Total Employment 1
HBK12 K-12 Enrollment 1
HBU University Enrollment 1
HBO Population .1168
HBO Retail Employment .8257
HBO Commercial Govt Employment .0408
HBO Other Employment .0137
NHBW NHBO Population .0575
NHBW NHBO Retail Employment .7593
NHBW NHBO Commercial Govt Employment .0846
NHBW NHBO Other Employment .1053
13
Trip Distribution

• Doubly-constrained Gravity model for the 6
  internal trip purposes E-I trips
• Friction factors
• - Initial factors from the old MINUTP model
• Fitting to Gamma function for smoothing the
  factors
• Adjustment of Gamma parameters to arrive at
  correct trip lengths by trip purpose
• Socioeconomic (or K) factors

14
Vehicle Occupancy

• Vehicle occupancy rates by trip purpose
• HBW ? 1.09 persons/vehicle
• HBK12 ? 1.96 persons/vehicle
• HBU ? 1.20 persons/vehicle
• HBO ? 1.83 persons/vehicle
• NHBW ? 1.17 persons/vehicle
• NHBO ? 1.75 persons/vehicle
• (Source The 2000 Knoxville Household Travel
  Behavior Study)

15
Time-of-Day

• AM peak (06300900) Midday peak
  (12001330) PM peak (15301800) and
  Off-peak
• TOD factors
• - Split the 24-hr trip table into tables by TOD
• - by trip purpose
• Directional factors
• - Convert trip tables in a production-attraction
  format to origin- destination tables
• - by trip purpose and by TOD
• Data Source the 2000 Knoxville Household Travel
  Behavior Study

16
External Trips

• 44 external stations
• E-E trips
• The 1994 O-D survey by Wilbur Smith Associates ?
  Most of E-E trip interchanges except for the
  interstate and US60 exchanges
• Kentucky Statewide Travel Model via critical
  link analysis ? E-E trip interchanges for the
  interstate and US60 external stations
• Modlin Equations ? A check of reasonableness
• E-I trips ADTs - E-E trips at each external
  station
• 27 stations of the 44 stations carry through
  trips

17
Traffic Assignment

• Time-of-day user equilibrium assignments
• - separate AM-peak, Midday-peak, PM-peak
  Off-peak assignments
• Improved free-flow speeds and varying capacities
  by TOD by reversible lanes
• Separate volume-delay functions (defaults)
• Unsignalized facilities ? ? 0.20 ? 10.0
• Signalized facilities ? ? 0.05 ? 10.0
• Feedback assignment

18
Feedback Loop
24-Hour Average Speed Gravity Model
TOD Trip Table Factoring
AM-peak Assignment
Midday-peak Assignment
PM-peak Assignment
Off-peak Assignment
24-Hour Weighted Average Congested Link Speed
19
CAL_REP

• Assignment post-processor written in GIS-dk and
  incorporated in the GUI
• Various error statistics by functional
  classification, volume group, screenline,
  cutline and specific corridors
• Error statistics
• total counts, average counts, counts standard
  deviation
• total loadings, average loadings, loadings
  standard deviation
• root mean square error
• mean error, error
• total counts VMT, total loadings VMT, VMT error

20
Assignment Results
21
CAL_REP Report

• Loading error -1
• Loading VMT error 0.9
• RMSE 25
• Interstate
• 0.03 ( error) -2.6 (VMT error)
• Urban Principal Arterial
• -1.3 -3.1
• Urban Minor Arterial
• -3.8 3.7

22
In Progress

• Model runs for future years and post-processing
• 2030 with committed projects only
• 2010 with committed plus Plan projects scheduled
  for completion by 2010
• 2020 with committed plus Plan projects scheduled
  for completion by 2020
• 2030 with all projects in the current LRP
• Multinet feature in GUI for efficient network
  data handling

23
THANK YOU!