danhlephuoc / best-scenario

Berlin Sumo Traffic (BeST) Scenario

Berlin Sumo Traffic (BeST) Scenario

best-mov.mp4

License and Attribution

All files belonging to this scenario definition are licensed under Creative Commons Attribution 4.0 International License .

The usage of this scenario has to be attributed by either providing a link to this repository, or by citing this reference:

Schrab, K., Protzmann, R., Radusch, I. (2023). A Large-Scale Traffic Scenario of Berlin for Evaluating Smart Mobility Applications. In: Nathanail, E.G., Gavanas, N., Adamos, G. (eds) Smart Energy for Smart Transport. CSUM 2022. Lecture Notes in Intelligent Transportation and Infrastructure. Springer, Cham. https://doi.org/10.1007/978-3-031-23721-8_24

24 Hours of Traffic in Berlin for SUMO and Eclipse MOSAIC

This simulation scenario provides motorized private transport traffic for over 24 hours for the whole city of Berlin. With about 2,25 million trips within an area of 800 km², this is the largest microscopic traffic simulation scenario we are currently aware of.

The scenario was made for Eclipse MOSAIC and thus requires it to run. It is, however, also possible to run it with Eclipse SUMO only. If you plan to test your own mobility solutions encoperating V2X technology or message exchange via LTE/5G, then Eclipse MOSAIC is your way to go. Here you can combine the traffic with communcation and application simulation, thus creating a holistic system solution on a large-scale.

The scenario is based on the MATSim Open Berlin Scenario ¹. We extracted traffic demand from this scenario and re-calibrated all routes to achieve a user equilibrium. More details on our creation process can be found in the provided reference and in the background section and the bottom of this file.

Characteristics

Some basic characteristics describing the scenario:

Characteristic	Number
Number of nodes	27 404
Number of edges	69 234
Number of junctions controlled by traffic signals	2 249
Number of trips	2 248 952
Average duration of each trip	805 s
Average distance of each trip	7,9 km
Overall mean speed compared to speed limits	0,71
Total number of teleports	2 786
Simulation duration on 3,4 GHz CPU (no GUI)	7 h

We furthermore compared the simulated counts on some certain streets against real data from Digitale Plattform Berlin:

Installation and Usage

1. Install Eclipse MOSAIC 22.0 ², e.g., by following this manual
2. Install Eclipse SUMO 1.11.0 ², e.g., from https://sumo.dlr.de/docs/Downloads.php
3. Clone this repository to an arbitrary folder.

   git clone https://github.com/mosaic-addons/best-scenario.git

4. To download the SUMO files for the scenario (~420 MB), execute the download_best_scenario.py³ script in /path/to/repository/scenario/sumo using Python 3.

   cd /path/to/repository/scenario/sumo
   py download_best_scenario.py

5. Go to the installation directory of Eclipse MOSAIC and type:

   mosaic.bat -c /path/to/repository/scenario/scenario_config.json -w 120 # Windows
   ./mosaic.sh -c /path/to/repository/scenario/scenario_config.json -w 120 # Linux

6. Be aware that completing this scenario requires several hours to complete. You can, however, reduce the simulation duration in the scenario_config.json.

In order to see a visualization of the traffic, simply edit the file etc/runtime.json in the Eclipse MOSAIC main directory. Replace SumoAmbassador with SumoGuiAmbassador and save the file. Then execute the scenario. Please note that using the visualization in sumo-gui slows down the simulation significantly due to its immense size.

...
  {
    "id": "sumo",
    "classname": "org.eclipse.mosaic.fed.sumo.ambassador.SumoGuiAmbassador",
    "configuration": "sumo_config.json",
    ...
  }
...

---

The scenario can also be used with SUMO only. Once you installed the scenario, you can execute it with SUMO directly:

sumo -c /path/to/repository/scenario/sumo/berlin.sumocfg

The MOSAIC scenario

This scenario is compatible with Eclipse MOSAIC. It is prepared to extends the traffic simulation in SUMO with communication and application simulation. You can easily enable and disable simulators in the bottom section of the scenario_config.json. Currently, only sumo and application is activated.

"federates": {
     "application": true,
     "cell": false,
     "sns": false,
     "sumo": true,    
     "output": false
}

In the mapping/mapping_config.json, you will find that 1% of all vehicles are equipped with a HelloWorldApp, which simply prints out the type of the vehicle in every simulation step. You can map our other Example Applications, or develop your own applications and map them onto a proportion of all vehicles.

{
  "prototypes":[
    {
      "name":"DefaultVehicle",
      "weight": 0.01,
      "applications":[ "org.eclipse.mosaic.app.tutorial.eventprocessing.sampling.HelloWorldApp" ]
    }
  ]
}

Furthermore, if you want to add communication between vehicles/their applications to the scenario, you can either activate sns for adhoc-communication, or cell for cell-based communication. Configuration files for both simulators, where you can configure delay times, paket loss, and other communication properties, can be found in sns/sns_config.json and cell/network.json files. To enable communication in applications, you furthermore need to activate the communication module accordingly in your application, and use it to send V2X messages. For more details on that, follow our tutorials.

Following an example for an application, which sends a Cooperative Awareness Message (CAM) via adhoc communication to its neighboring vehicles:

public class V2xApp
   extends AbstractApplication<VehicleOperatingSystem>
   implements CommunicationApplication, VehicleApplication {

   public void onStartup() {
     getAdhocModule().enable();
   }

   public void onVehicleUpdated() {
     getAdhocModule().sendCam();
   }

   public void onMessageReceived(ReceivedV2xMessage msg) {
     if (msg.getMessage() instanceof Cam) {
       String senderId = msg.getMessage().getRouting().getSource().getSourceName();
       GeoPoint otherPos = ((Cam)msg.getMessage()).getPosition();
       // todo
     }
   }
}

Footnotes

1. D. Ziemke, I. Kaddoura, K. Nagel; The MATSim Open Berlin Scenario: A multimodal agent-based transport simulation scenario based on synthetic demand modeling and Open Data; Procedia Computer Science, Volume 151, 2019, 870-877 ↩

2. We calibrated and tested the BeST scenario using the mentioned versions of SUMO and MOSAIC. The scenario may still work with newer versions, but we cannot guarantee that the same results will be created. ↩ ↩²

3. The download executed by this script will be counted for statistical purposes on www.dcaiti.tu-berlin.de. To disable recording the download, you can set the field record in the download_best_scenario.py to False. Details about tracking on that site can be found at https://www.dcaiti.tu-berlin.de/contact/imprint ↩