Skip to content

Autonomous Vehicles Intelligence Analysis

License

Notifications You must be signed in to change notification settings

amodeus-science/amod

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

amodeus.amod Build Status

This repository allows to run an autonomous mobility-on-demand scenario using the amodeus library.

Try it, orchestrate your own fleet of amod-taxis! Watch a visualization of a traffic simulation in San Francisco generated using this repository.

p1t1

p1t4

p1t3

p1t2

Admins

AMoDeus is jointly maintained and further developed by the Admins and Code Owners Christian Fluri (ETH Zürich), Joel Gächter (ETH Zürich), Sebastian Hörl (ETH Zürich), Claudio Ruch, Jan Hakenberg, ChengQi Lu (TU Berlin), and Marc Albert (nuTonomy).

Please let us know if you'd like to contribute!

First steps in the amod repository

Prerequisites

  • You may work on a Linux, Mac or Windows OS with a set of different possible IDEs. The combination Ubuntu, Java 8, Eclipse has worked well.
  • Install Java SE Development Kit (version 8, or above)
  • Install Apache Maven
  • Install IDE (ideally Eclipse Oxygen or Photon)
  • Install GLPK and GLPK for Java (Ensure you install compatible versions, e.g. here)
    • Prerequisites are: GCC, Libtool, Swig and Subversion
  • Install Git and connect to GitHub with SSH

The code format of the amod repository is specified in the amodeus profile that you can import from amodeus-code-style.xml.

Getting Started

Follow these step-by-step instructions or the video to set up, prepare, and run your first simulation. You can get a sample simulation scenario at https://www.amodeus.science/

  1. Clone amod
  2. Import to eclipse as existing maven project (Package Explorer->Import) using the pom.xml in the top folder of this repository.
  3. Set up Run Configurations for: (ScenarioPreparer; ScenarioServer; ScenarioViewer), chose the Working Directory to be the top Simulation Folder directory.
  4. Adjust the simulation settings in the 3 config files: av.xml for av fleet values (e.g. number vehicles), AmodeusOptions.properties for AMoDeus settings (e.g. max number of people) and config.xml for Matsim settings (e.g. output directory).
  5. Add JAVA VM arguments if necessary, e.g., -Xmx10000m to run with 10 GB of RAM and -Dmatsim.preferLocalDtds=true to prefer the local Dtds.
  6. Run the ScenarioPreparer as a Java application: wait until termination
  7. Run the ScenarioServer as a Java application: the simulation should run
  8. Run the ScenarioViewer as a Java application: the visualization of the scenario should open in a separate window

Gallery

usecase_amodeus

Zurich

p1t5

San Francisco

San Francisco

San Francisco

String Interface

TODO improve and adapt documentation, de-dockerize Run docker-compose up to run the San Fransisco simulation. This will run two services, aido-host and aido-guest, which will communicate over port 9382.

The protocol is specified here.


ethz300