L4 AIMSUN Overview and Algorithms

8/12/2019 L4 AIMSUN Overview and Algorithms

1/45

AIMSUN

Advanced Interactive Microscopic

Simulator for Urban and Non-urban

Networks

Adopted from Clara Fang/ Ondrej Pribyl


2/45

2

Outline

Introduction

Practical Applications

Capabilities

Simulation Requirements Simulation Inputs

Simulation Outputs

Function Limitations


3/45

3

Introduction

Developer Transportation Simulation Systems (TSS), Barcelona, Spain

www.aimsun.com

AIMSUN Ver 4 is integrated with GETRAM SimulationEnvironment

Generic Environment for TRaff ic Analysis and Model ing

Traffic Network Graphic Editor (TEDI)

AIMSUN

AIMSUN 3D


4/45

4

GETRAM

External Applications

ShortestRoutes

Component

Network

Database

TEDI

GraphicalEditor

Costs

Routes

GETRAM

AIMSUN

-

User Interface

AIMSUN2

KernelGETRAM

Extensions

Simulated

Data

Control &

Management

Actions

EMME/2

TRANSYT

SCATSInterfaces


5/45

7

Capabilities

Urban networks, freeways, highways, ring roads,interchanges, roundabouts, arterials and anycombination of them

Public transportation

Traffic incidents Vehicle types

cars,buses, trucks, trains or user-defined

Fixed vs.Dynamic route choice models

Interfaces EMME/2

TRANSYT

3D visualization


6/45

8

Capabilities

Traffic Control and Management

Traffic signals (NEMA controllers)

fixed-time, semi-actuated, fully-actuated

Adaptive control

Signs

Ramp metering

Green time, flow, delay

VMS

Different message and Starting time

All kinds of detectors


7/45

9

Simulation Requirements

(1) Network Geometry Vehicles data

(2) Traffic Demand

(3) Traffic Control Plan

(4) Public Transportation (optional)

(5) GETRAM Extensions (optional)

Modelling parameters (Default values areprovided)

Scenario


8/45

10

Simulation Inputs

(1) Network Geometry

Sections (links)

length, width, number of lanes, speed limits, grade, etc.

Nodes (Junctions and Joins)

turning movements for junctions

Centroids

traffic source, sink or both

Vehicles type, size, vehicle characteristics

Detectors, VMS, etc.

Map of the area (optional)


9/45

11

TEDI: a user friendly graphic interface for building models


10/45

12

Network Model


11/45

13

Examples of Sections and Polysections


12/45

14

Section Properties


13/45

15

Examples of Joins


14/45

17

Intersection - Turning Movements


15/45

18

Centroids


16/45

19

Vehicle Types

Transfer between Library and Model


17/45

20

Importing backgrounds as .jpg, .bmp, .tif ,...files


18/45

21

Simulation Inputs

(2) Traffic Demand

Result Based - Traffic Flows & Turning

Proportions

generated at origin centroids and input into the network

through the sections connected to the centroid distributed around the network in accordance to the

turning proportions defined in each section of the

network.

Route Based - O/D matrix and Shortest Paths

generated at origin centroids and input into the networkthrough the sections connected to the centroid

distributed following shortest paths from input section

to destination centroid.


19/45

22

Traffic Flows and Turning Proportions

Define State


20/45

23

Simulation Inputs

Signal

Signal types

Signal groups (turning movements are grouped)

Phases sequences and associated signals groups

Duration of each phase Offset

Actuated parameters

Unsignalized

Priority rules - Yield and Stop sign Parameters that affect the Gap-acceptance model

Ramp metering

Control parameters (green time, flow or delay time)


21/45

24

OD Matrix


22/45

25

Traffic Generation

Arrival Distributions

Exponential

Route based modeling

Uniform

Result based modeling

Normal

Constant

Other Arrival Models ASAP (as soon as possible)

External (GETRAM Extensions)


23/45

26

Simulation Inputs(3) Traffic Control

Signal

Signal types

Signal groups (turning movements are grouped)

Phases sequences and associated signals groups

Duration of each phase Offset

Actuated parameters

Unsignalized

Priority rules - Yield and Stop sign Parameters that affect the Gap-acceptance model

Ramp metering

Control parameters (green time, flow or delay time)


24/45

27

Typical signal plan


25/45

28

Uncontrolled

Fixed

External Actuated

SCATS

Traffic Signal Control Types


26/45

29

Traffic Signal Control Plan


27/45

30

Simulation Inputs(4) Public Transport (Optional)

Public transport lines

Routs

Reserved lanes

Bus stops

Vehicle type

Timetables

Departure frequency

Fixed schedule

PT Plan


28/45

31

Simulation Inputs

(5) GETRAM Extensions (Optional)

Application Programming Interface (API)

Develop external applications (traffic control

systems) Access to the statistical data produced by

simulated detectors, VMS or ramp metering

Keep track of a guided vehicle throughout the

network and directly control a vehicle movement Programming in C/C++,or using Python scripting

language


29/45

32

GETRAM extensions

Principle of data exchange


30/45

33

Simulation Inputs(6) Modelling parameters

Global Reaction time, queue up speed and queue leaving

speed, etc.

Car-following model

Maximum number of vehicles, maximum distance, etc.

Lane changing model Percent overtaken, percent recover, etc.

Local Speed limit, turning speed, visibility distance at

intersection, distance zone, etc.

Vehicle Attributes


31/45

34

Simulation Algorithms

Vehicle Arrivals

Vehicle Attributes

Global Simulation Parameters

Car-Following Lane Changing

Gap Acceptance


32/45

35

Vehicle Arrivals

User may select among the following models: Exponential

Uniform

(Truncated) Normal

ASAP Constant

External Source


33/45

36

Vehicle Attributes

Length

Width (considered for graphics only)

Maximum Desired Speed

Maximum Acceleration

Normal Deceleration

Maximum Deceleration

Speed Limit Acceptance

Minimum spacing


34/45

37

Global Simulation Parameters

Turning Speed Effects of Grade on Vehicle Performance

Drivers reaction time - also the simulation timestep - affects capacity!

Reaction time when vehicle is stoppedaffectsqueuing

Speed to join the queue

Speed to depart from a queue


35/45

38

Models of vehicle movements

The vehicle movements is computed based onparticular sub-models such as

Car-following model

Lane changing

The vehicles are aiming to get to the desired speed

But are constrained by environment

Adjacent vehicles, speed limits, signal light,


36/45

39

Principle

In every simulation step (time based triggering)are the parameters recomputed according tofollowing principle:

if (it is necessary to change lanes) then Apply Lane-Changing Model

endifif (the vehicle has not changed lanes) then Apply Car-Following Model

endif


37/45

40

Car-Following

Based on Gipps Model, which is a function of: Type of Driver

Geometry of the Section

Uses a 2-lanes car-following model to consider the effects

of adjacent lanes as a function of: Area to be considered

Number of vehicles in area

C t ti f l ti d d l ti


38/45

41

Computation of acceleration and deceleration (SourceAIMSUN Manual)


39/45

42

Lane Changing

Uses the Gipps Lane Changing Model It is based on:

Necessity of Lane Change

Desirability

Feasibility

Algorithm asks each vehicle at every update: Isit necessary, desirable, feasible to change lanes?


40/45

43

Gap-Acceptance

Used to answer the question: Is it feasible to changelanes?

Evaluates gap to both upstream and downstream vehicles.

There are segemetn sections defined in order to achieve

more representative behavior


41/45

44

Gap acceptance model (Source AIMSUN Manual)


42/45

45

Simulation Outputs

Statistical measures

In network, O/D matrix, stream, section,

turning, type of vehicles level

Mean Flow

Density

Mean Speed and Harmonic Mean Speed

Travel Time

Delay Time

Stop Time and Number of Stops

Queue Length (Mean and Maximum)

Total Travel Length

Fuel Consumed and Pollution Emitted


43/45

46

Simulation Outputs

Storage of simulation outputs ASCII files

Database (e.g. Microsoft Access)

User- defined time interval

Multiple runs Record simulation

Comparisons of two sets of time series data(Validation)

Provides charts and statistical indicators e.g., flows measured at a detector at each interval

(5 minutes) vs. field data


44/45

47

Function Limitations

No signal optimization

No control delay output

No signal phase switch information output


45/45

48

Thank you !

Questions & Comments?

L4 AIMSUN Overview and Algorithms

Documents