The server that hosted the original ngsim-community.org Web site suffered an irreversible failure that resulted in the loss of the on-line collaboration tools. However, many documentation and data download files were recovered, and are available again via the links below.

The Next Generation Simulation (NGSIM) program was initiated by the United States Department of Transportation (US DOT) Federal Highway Administration (FHWA) in the early 2000's.  The program developed a core of open behavioral algorithms in support of traffic simulation with a primary focus on microscopic modeling, and collected high-quality primary traffic and trajectory data intended to support the research and testing of the new algorithms. The NGSIM program has actively engaged traffic simulation vendors to accelerate the inclusion of advanced or improved algorithms in the commercial models used across the world. 

 


NGSIM File Downloads


NGSIM Basics

NGSIM Data Format Plan

This report presents the Data Format Plan developed as part of the NGSIM program. The plan defines the structure, documentation, and transfer requirements for data that will be collected for estimation, calibration, and validation of core behavioral algorithms. The formats developed in the plan are based on existing formats that are relevant to NGSIM. The plan specifies a NGSIM data model, a data dictionary based on P1489-1999 data standard, and formats for metadata and data exchange.

NGSIM Data Plan

This document presents a High-Level Data Plan for the collection of datasets to be used in developing and validating core behavioral algorithms under the NGSIM program. This High-Level Plan focuses on traffic data that are most useful to microscopic simulation model development and validation. The NGSIM team reviewed existing traffic datasets in terms of level of detail and coverage. Using this framework, the team has divided its data survey and assessment efforts into vehicle trajectory, instrumented vehicle, and detector data collection efforts.

Arterial Lane Selection Model

This report describes models for lane selection on congested arterial corridors, taking into account lane-choice prepositioning based on path plan considerations of the driver, target-lane choice, accepting the gaps to make a lane change towards the target lane, and execution of the lane change to the accepted gap. This report describes the structure, estimation results, implementation details and operation validation results.

NGSIM Verification and Validation Plan

This report presents a High-Level Plan for the Verification and Validation (V&V) of algorithms developed as part of the NGSIM program. The high-level plan presents definitions, describes algorithm interactions, describes the various stages of the V&V process, and provides an example of how V&V will be carried out. It also specifies roles for a researcher, an independent verifier, and an expert panel. The first two sections of this report provide an introduction to the verification and validation process and how the interaction among behavioral algorithms can affect algorithm validity.

NGSIM Core Algorithms Assessment

This report categorizes, assesses, and evaluates microsimulation behavioral models to develop a comprehensive historical view of the state-of-the-art. It also provides a review of the state-of-the-practice by examining how and where influencing factors are implemented in the most widely-used simulation systems. Earlier NGSIM reports produced a prioritization of core algorithm categories and influencing factors that are most connected to the simulation scenarios offered by the stakeholder groups. This report expands upon these general research directions by first reviewing behavioral model research for each of the core algorithm areas and then providing an evaluation of each algorithmic area in terms of: a) research and documentation; b) sensitivity and extensibility to influencing factors; c) computational complexity; and d) validation, estimation, and calibration.

NGSIM Core Algorithm Categories

This report focuses on identifying the core traveler behavioral algorithm categories in traffic simulation and prioritizing those core algorithm categories in terms of areas of research required to improve the quality and performance of simulation tools. General recommendations for NGSIM include: a) improving the simulation of operational driving behaviors to include new influencing factors;† b) developing simulation algorithms representing tactical route execution behaviors;† and, c) improving the simulation of strategic route modification behaviors based on response to information.

Cooperative Lane Changing and Forced Merging Model

This report describes the detailed structure, estimation results, implementation details, verification, and operation validation results of the models for cooperative lane changing and forced merging. The model presented in this report is a combined merging model that explicitly considers cooperation and competition between target lane drivers, and integrates normal, cooperative, and forced lane changing considerations of drivers in a single framework.

Open Source MITSIMLab

"NGSIM is using MITSIMLab to validate some of the algorithms developed under NGSIM.† In order to facilitate verification of the algorithm validation effort MIT (a member of the NGSIM team) has made MITSIMLab available to the simulation community.† MITSIMLab is available with an open source license by clicking the title above.† MITSIMLab Executables, source code, support information and user manual, and a discussion group are available at the website.

Freeway Lane Selection Algorithm

This report summarizes the NGSIM effort to estimate, calibrate, and validate a freeway lane selection model. The model developed in this study is a generalized lane changing model that explicitly incorporates the selection of a target lane.

Model for Cooperative Lane Changing and Forced Merging with Acceleration

This report describes the structure, estimation results, implementation details and operation validation results of the integrated lane changing and acceleration models for freeway merging. This is an extension of the Cooperative Lane Changing and Forced Merging model developed previously as part of the NGSIM program. The model presented in this report is a combined merging model that integrates the lane changing and acceleration decisions of the driver for normal, cooperative, and forced merging.

History of Transportation Simulation Models Diagram

This diagram summarizes the major research threads and the core contributions in traffic microsimulation modeling. Four major research threads are presented in the diagram including Route Modification, Lane Changing, Gap Acceptance and Car-Following.

NGSIM Framework

Simulation framework defining the focus of NGSIM - presented by the Principal Investigator

NGSIM Work Scope

Cambridge Systematics, Inc. scope of work for the NGSIM program.

NGSIM Program Information

Detailed background and introduction to the NGSIM program.

FHWA Overview

NGSIM Overview - presented by the FHWA

NGSIM Stakeholder Presentations at TRB 2007


The TRB NGSIM stakeholder meeting took place on Sunday, January 21, 2007 at the Marriott Hotel in the Coolidge room from 8:30 AM to noon.† The agenda was:

  • 8:30-8:45 AM -- Summary of NGSIM activities in 2006 -- Vassili Alexiadis, Cambridge Systematics, Inc.
  • 8:45-9:00 AM -- The NGSIM Freeway Lane Selection algorithm in CORSIM -- Li Zhang, Mississippi State University.
  • 9:00-9:45 AM -- Cooperative/Forced Freeway Merging Algorithm -- MIT presented estimation/validation results for this algorithm.† MIT also presented on the plan for incorporating acceleration in this algorithm.
  • 9:45-10:30 AM -- Oversaturated Freeway Flow Algorithm -- The University of California presented estimation/validation results for this algorithm.† †
  • 10:30-11:15 AM -- Arterial Lane Selection Algorithm -- MIT presented estimation/validation results for this algorithm.
  • 11:15-11:30 AM -- An upcoming NGSIM algorithm -- Michael Kyte, University of Idaho
  • 11:30 AM-noon -- The future of NGSIM -- The FHWA presented on the future of NGSIM.

Note: This file is the collection ('.zip file) of the following:
  • The future of NGSIM
  • Arterial Lane Selection Algorithm
  • Oversaturated Freeway Flow Algorithm
  • Cooperative/Forced Freeway Merging Algorithm
  • The NGSIM Freeway Lane Selection algorithm in CORSIM
  • Summary of NGSIM Activities in 2006

The file is 108 MB in size.

The future of NGSIM

FHWA gave this presentation on the future of NGSIM

Algorithms

Arterial Lane Selection Algorithm

MIT gave this presentation on the estimation/validation results for this algorithm

Oversaturated Freeway Flow Algorithm

The University of California gave this presentation on the estimation/validation results for this algorithm

Cooperative/Forced Freeway Merging Algorithm

MIT gave this presentation on the estimation/validation results for this algorithm. MIT also gave a presentation on the plan for incorporating acceleration in this algorithm

The NGSIM Freeway Lane Selection algorithm in CORSIM

Li Zhang, Mississippi State University.

Summary of NGSIM Activities in 2006

Annual Workshop on Simulation - TRB 2006 Presentations

The TRB Annual Workshop on Simulation focused on NGSIM this year.† This workshop took place on Sunday, January 22, 2006 from 1:30 to 5:00 PM.† In addition to presentations on the overview and future steps of NGSIM, presentations on data and algorithm were given by NGSIM team members, software developers and NGSIM data users.

  1. Overview on Traffic Analysis Tools and NGSIM programs
  2. Data

    • Cross-cutting look at vehicle trajectory data from different sites
    • Vehicle trajectory processing software - NG-VIDEO demonstration
    • Using NGSIM data to comparing HCM with other capacity/speed measures

  3. Algorithms

    • Results from the commercial validation of the first NGSIM algorithm – Freeway lane selection
      • AIMSUN, PARAMICS, VISSIM

    • A new algorithm for cooperative/forced freeway merging †† †
    • Research plan for a new NGSIM algorithm – Oversaturated freeway flow

NGSIM Stakeholder Presentations at TRB 2006

The NGSIM Stakeholder Meeting on Sunday, January 22, 2006 focused on behavioral algorithm research.


  • Dr. Tomer Toledo presented on the status of ongoing research for modeling cooperative/forced merging at freeway onramps.† The presentation was primarily focused on the preliminary estimation results of the algorithm.

  • Dr. Alex Skabardonis of University of California presented the draft research plan for a new Oversaturated Freeway Flow algorithm.† Dr. Skabardonis described an improved model for oversaturated freeway flow that focuses on both car-following and lane-changing behaviors during congested conditions.

  • Lastly, Dr. Toledo presented the draft research plan for a new Arterial Lane Selection algorithm.† The lane selection model will incorporate earlier NGSIM research on lane changing and develop an improved gap selection methodology.

Second NGSIM Stakeholder Workshop Presentations

This ZIP file contains the presentations from the second annual NGSIM stakeholder workshop, held on January 10, 2004 in Washington D.C. The NGSIM team presented updates on the various tasks including Validation Framework and Discussion, Algorithm Prioritization and Assessment, Algorithm Development Process, and Data Collection Plan and Data Formats.

TRB - Workshop on Traffic Simulation Models

This ZIP file contains the overview, update and progress report of the NGSIM program to the transportation community at a workshop on Traffic Simulation Models. The workshop was held on the Sunday, Jan 11 at the 2004 TRB Annual Meeting, Washington, D.C.

NGSIM Stakeholder Meeting Agenda

"Agenda for the NGSIM Stakeholder meeting on Saturday, January 10 at the Shoreham Hotel - Governors Room in Washington, DC.

March 2003 Workshop Press Release

March 2003 Workshop Press Release.

Scenario Example 3

Material used in the Users and Modelers breakout sessions

Scenario Example 2

Material used in the Users and Modelers breakout sessions

Scenario Example 1

Material used in the Users and Modelers breakout sessions

Scenario Form

Material used in the Users and modelers breakout sessions

Scenario Definitions and Instructions

Material used in the Users and Modelers breakout sessions

Breakout instructions

Instructions for breakout sessions - presented by the Investigator

NGSIM General

NGSIM approach and workshop objectives - presented by the Principal Investigator

TRB Workshop slides

"Slides for the NGSIM Workshop presentation during the 2003 TRB in Washington, D.C.

Stakeholder Workshop Agenda

This is the Agenda for the March 24 Stakeholders workshop

Traffic Simulation Workshop Agenda

The TRB Committees on Traffic Flow Theory and Characteristics and on Highway Capacity and Quality of Service, are co-sponsoring this workshop to address recent advances in the development and application of traffic simulation models in the design, evaluation and operation of modern transportation facilities.

Modeler Session Notes

These are the notes from the Modeler Session held during the March 2003 Stakeholder meeting

User Session Notes

These are the notes from the User Session held during the March 2003 Stakeholder meeting

User Session Tuesday

Introduction to the User stakeholder group in Tuesday's breakout sessions - presented by the User stakeholder group Chair

User Session Monday

Introduction to the User stakeholder group in Monday's breakout sessions - presented by the User stakeholder group Chair.

User's Plenary

User's group goals and membership - presented by the user stakeholder group Chairperson

Developer Session Notes

These are the notes from the Developers session held during the March 2003 Stakeholders workshop

Developer Session Tuesday

Introductions to the Developer stakeholder group in Tuesday's breakout sessions - presented by the Developer stakeholder chair

Developer Session Monday

Introduction to the Developer stakeholder group in Monday's breakout sessions - presented by the developer stakeholder group chair

Developer's Plenary

Developer's group goals and membership - presented by the developers stakeholder group Chairperson

Modeler Session Tuesday

Introduction to the Modeler stakeholder group in Tuesday's breakout sessions - presented by the Modeler stakeholder group Chair

Modeler Session Monday

Introduction to the Modeler stakeholder group in Monday's breakout sessions - presented by the modeler stakeholder group Chair.

Modeler's Plenary

Modeler's group goals and membership - presented by the modelers stakeholder group Chairperson

NGSIM article in Public Roads


An article on NGSIM was published in the January/February issue of the FHWA's Public Roads magazine.† The article details NGSIM objectives, quotes from various stakeholders, and a summary of NGSIM and datasets. †


The link to the article is here

NGSIM article in R&T Transporter


An article on NGSIM appeared in the June issue of the U.S. DOT's Research and Technology Transporter. The article provides a summary of the history of the NGSIM program, and describes the Freeway Lane Selection algorithm, the three NGSIM algorithms currently under development, and the NGSIM datasets.

The link to the article is here

ITS Berkeley Online Magazine


David Downs, ""Next Generation Simulation"", NewsBITS, The ITS Berkeley Online Magazine Fall 2005: Volume 1, Number 1

ITE NGSIM Article (ITE online membership required)


Vassili Alexiadis, James Colyar, John Halkias, Rob Hranac, and Gene McHale, ""Next Generation Simulation Program"", ITE Journal, August 2004.

/Downloads/Stakeholders/ITE_NGSIM.pdf

NGSIM Videos

Peachtree

Data Analysis

Peachtree Main Data

Peachtree Meta Data

Raw Video

"Raw video, 12:45 - 1:00 PM, cameras 1 thru 2"


"Raw video, 12:45 - 1:00 PM, cameras 3 thru 5"


"Raw video, 12:45 - 1:00 PM, camera 6"


"Raw video, 12:45 - 1:00 PM, camera 7"


"Raw video, 12:45 - 1:00 PM, camera 8"


"Raw video, 4:00 - 4:15 PM, camera 3"


"Raw video, 4:00 - 4:15 PM, camera 8"


"Raw video, 4:00 - 4:15 PM, cameras 1 thru 2"


"Raw video, 4:00 - 4:15 PM, cameras 4 thru 5"


"Raw video, 4:00 - 4:15 PM, cameras 6 thru 7"


Processed Video

"Processed Video 12:45 - 1:00 PM, cameras 1 thru 4"


"Processed Video 12:45 - 1:00 PM, cameras 5 thru 8"


"Processed Video 4 - 4:15 PM, cameras 1 thru 4"


"Processed Video 4 - 4:15 PM, cameras 5 thru 8"


Lankershim

Data Analysis Report


Main Data


Raw Video

"Raw Video, 8:30 AM - 8:45 AM, cameras 1 thru 3"


"Raw Video, 8:30 AM - 8:45 AM, cameras 4 thru 5"


"Raw Video, 8:45 AM - 9:00 AM, cameras 1 thru 3"


"Raw Video, 8:45 AM - 9:00 AM, cameras 4 thru 5"


Processed Video

"Processed Video 8:30 AM - 8:45 AM, cameras 1 thru 3"


"Processed Video 8:30 AM - 8:45 AM, cameras 4 thru 5"


"Processed Video 8:45 AM - 9:00 AM, cameras 1 thru 3"


"Processed Video 8:45 AM - 9:00 AM, cameras 4 thru 5"


US-101

Main Data


Data Analysis Report


Raw Video

"Raw video, 7:50 AM - 8:05 AM, cameras 1 thru 4"


"Raw video, 7:50 AM - 8:05 AM, cameras 5 thru 8"


"Raw video, 8:05 AM - 8:20 AM, cameras 1 thru 4"


"Raw video, 8:05 AM - 8:20 AM, cameras 5 thru 8"


"Raw video, 8:05 AM - 8:20 AM, cameras 5 thru 8"


"Raw video, 8:20 AM - 8:35 AM, cameras 1 thru 4"


"Raw video, 8:20 AM - 8:35 AM, cameras 5 thru 8"


Processed Video

"processed video 7:50 AM - 8:05 AM, cameras 1 thru 4"


"processed video 7:50 AM - 8:05 AM, cameras 5 thru 8"


"processed video 8:05 AM - 8:20 AM, cameras 1 thru 4"


"processed video 8:05 AM - 8:20 AM, cameras 5 thru 8"


"processed video 8:20 AM - 8:35 AM, cameras 1 thru 4"


"processed video 8:20 AM - 8:35 AM, cameras 5 thru 8"


I-80

Meta Data


Main Data


Data Analysis Report


Raw Video

"Raw video, 4:00 PM - 4:15 PM, cameras 1 thru 4"


"Raw video, 4:00 PM - 4:15 PM, cameras 5 thru 7"


"Raw video, 5:00 PM - 5:15 PM, cameras 1 thru 4"


"Raw video, 5:00 PM - 5:15 PM, cameras 5 thru 7"


"Raw video, 5:15 PM - 5:30 PM, cameras 1 thru 4"


"Raw video, 5:15 PM - 5:30 PM, cameras 5 thru 7"


Processed Video

"Processed video 4:00 PM - 4:15 PM, cameras 1 thru 4"


"Processed video 4:00 PM - 4:15 PM, cameras 5 thru 7"


"Processed video 5:00 PM - 5:15 PM, cameras 1 thru 4"


"Processed video 5:00 PM - 5:15 PM, cameras 5 thru 7"


"Processed video 5:15 PM - 5:30 PM, cameras 1 thru 2"


"Processed video 5:15 PM - 5:30 PM, cameras 1 thru 4"


"Processed video 5:15 PM - 5:30 PM, cameras 3 thru 4"


"Processed video 5:15 PM - 5:30 PM, cameras 5 thru 7"


Miscellaneous

Data Analysis Report


Main Data


Raw video


Processed Video


Aerial Photos


The SAVME Data Set


Berkely Data Set

Sample data from University of California at Berkeley Highway Laboratory.

The Atlas Data Set

Sample dataset from ATLAS/University of Arizona.

NGSIM-Video User's Manual

The User's Manual provides additional information on the skills, video formats and resources necessary to operate the program. The manual provides: (1) NGSIM-VIDEO installation steps, (2) guidelines to ensure collected videos meet the image requirements for NGSIM-VIDEO processing, (3) preprocessing of the video for NGSIM-VIDEO detection and tracking, (4) video processing through the NGSIM-VIDEO program to obtain vehicle trajectories, and (5) database checking to ensure that the trajectories are accurate.

Oversaturated Freeway Flow Model

"Existing simulation models have difficulty in accurately modeling oversaturated traffic conditions on freeways. This report describes a behavioral algorithm for oversaturated freeway flow that can be used in microscopic simulation models. The algorithm is an integrated car-following and lane-changing modeling framework that is consistent with the kinematic wave theory. The algorithm can explicitly model mandatory and discretionary lane changing, including cooperation during lane changing. As an extension of the algorithm, a new on-ramp merging model was developed and incorporated into the model. Also, the proposed algorithm accounts for the relaxation process following lane changing. The model includes a limited number of parameters that can be readily measured in the field. The report includes the algorithm description, analysis results and computer code.

NGSIM Video source and program installer

"Raw video, 8:20 AM - 8:35 AM, cameras 1 thru 4"


"Raw video, 8:20 AM - 8:35 AM, cameras 5 thru 8"