Next Generation Simulation NGSIM Program

1
Next Generation Simulation (NGSIM) Program

• A Status Update

presented by Vassili Alexiadis Cambridge
Systematics, Inc. January 11, 2004
2
Task Overview
Demand
Supply
September 03
December 03
Algorithm Prioritization
Algorithm Assessment
Market Assessment
October 03
January 04
January 04
High Level Validation Plan
High Level Algorithm Plan
High Level Data Plan
Planning
Data Formats
Prototyping Efforts
Project Infrastructure
Website
3
Accomplishments to Date

• Consultation with stakeholders Workshop at
  University of Maryland (March 2003) Conference
  calls
• Established NGSIM web site and discussion forum
• Identification and prioritization of core
  algorithm categories
• High-level plan for algorithms
• High-level plan for datasets
• Data format plan
• Sample vehicle trajectory datasets

4
Identification and Prioritization of Core
Algorithm Categories

• Objective
• Identify and prioritize the core traveler
  behavioral algorithm categories in terms of areas
  of research required to improve the quality and
  performance of simulation tools
• Process
• Identified a structure for traffic simulation
  core components
• Divided core components into behavioral
  algorithms and influencing factors
• Developed scenario requirements methodology
• Using scenarios, stakeholders identified
  priorities for NGSIM

5
Core Algorithm Categories Recommendations
Time to Make/Execute Decision

• Focus on route execution algorithms to make a big
  impact in addressing gaps
• Make a quick impact with modeling basics by
  considering a broader range of influencing
  factors in operational driving algorithms
• Improve the simulation of strategic en-route
  traveler behaviors based on response to
  information

Pre-Trip
Macroscopic
Strategic En-Route
30 seconds
TacticalRoute Execution
NGSIMFocus
5 seconds
OperationalDriving
Nanoscopic
VehicleControl
6
High-Level Plan for Algorithms

• Objective
• Categorize, assess and evaluate microsimulation
  behavioral algorithms from both theoretical and
  implementation perspectives
• Process
• Identified evaluation criteria
• Conducted detailed evaluation of research on
  traveler behavior - Evaluated algorithms
• Conducted detailed survey of existing influencing
  factors
• Developed process for data collection, and
  development and validation of algorithms
• Recommendations

7
High-Level Plan for Algorithms - Recommendations
3
Route Modification
Strategic
Parking
Emergency Response
Cooperative Behaviors
Tactical
1
Merging
BehavioralAlgorithm Categories
Mixed-Mode
Lane Changing
Gap Acceptance
Acceleration
Operational
Queue Discharge
Transit
Pedestrian
Low
High
Research Documentation
8
Recommendations

• Identify specific models to be implemented as
  algorithms and validated against collected data
  sets
• For models where there is significant existing
  research and documentation, the key issue is
  acceptance by the modelers and users community
• Develop an algorithm (software implementation of
  a model) that can be validated against a
  collected data set within the NGSIM framework
• For models where there is a clear theoretical
  gap, either in the fundamental model of in the
  consideration of influencing factors,
• Adopt a promising modeling approach and
  investigate enhancing the model
• Develop an algorithm based on the new model and
  evaluate it against a collected data set

9
High-Level Plan for Datasets

• Objective
• Develop a high-level plan for the collection of
  data sets to be used in developing and validating
  core simulation algorithms under the NGSIM
  program
• Process
• Reviewed existing traffic data sets in terms of
  level of detail and coverage
• Conducted three data surveys and assessment

10
HLP for Datasets Recommendations

• Focus on collecting vehicle trajectory data and
  related aggregate data
• Engage in two data collection efforts for vehicle
  trajectory data on freeway and signalized
  arterial environments
• Augment with associated aggregate data
• Partner with existing data collection efforts
• Utilize existing technology
• Collect original data
• Design data collection plans so that the data
  needs of multiple scenarios can be met 
• Assemble relevant and easily accessible aggregate
  data and place these datasets in the NGSIM data
  repository

11
Data Format Plan

• Objective
• Develop documentation requirements for NGSIM data
  formats
• Process
• Review existing data definitions and formats
• Discuss with stakeholders
• Recommendations
• Metadata
• Data dictionary
• Data model
• Data exchange format

12
Coming Soon

• High-level plan for validation
• Prototype dataset
• Prototype algorithms
• Task orders for
• Data collection
• Algorithm development and validation