Simulations#

The Simulations page contains the list of test case simulations you can execute. It serves as the central place from which you can create new simulations, as well as start and stop your simulations. A tutorial video on simulations creation and management can be found here.

Creating a new simulation top#

Please refer to the Running Simulator document for detailed instructions on creating a new simulation.

The following pages denote the pieces of a new simulation configuration.

General top#

  • Simulation Name - Enter a name for your simulation
  • Description (optional) - Enter a description for your simulation
  • Tags (optional) - Specify tags for your simulation, which can be used to search through or filter simulations later
  • Select Cluster - Choose the simulation cluster which will actually run the simulator executable and execute the simulation. If you select a cloud cluster, the following three toggles can not be changed.
  • Create test report - Enable to generate a test report after execution.
  • Headless mode - Enable to run the simulator without rendering visually in the executable
  • Interactive mode - Enable to run the simulator interactively, with the ability to visualize sensors, pause, etc. within the executable

Test case top#

Select a runtime template:

  • Random Traffic: A random simulation where starting conditions are defined here in the user interface
  • Visual scenario editor: Run a scenario created by the Visual Scenario Editor
  • Python API: Directly input the Python API test case script to automatically execute from the user interface. If you have selected a local cluster, you will still need to bring up your Autopilot separately.
  • API Only: The simulator is simply started in API Only mode, and you are expected to use a runner to execute test cases (Python API, Visual Scenario Editor)

Curently, only the Python API runtime template is available if you have selected a cloud cluster.

Runtime Template: Random Traffic#

  • Map: Select the map for the simulation
  • Vehicle: Select one or more ego vehicles
  • Simulation Date: Set the date within simulation
  • Time of Day: Set the time of day
  • Rain: [0-1] set how much rain should fall
  • Fog: [0-1] set how thick fog should be
  • Wetness: [0-1] set how wet the road surfaces should be
  • Cloudiness: [0-1] set how much cloud cover there should be
  • Random Traffic: Enable for non-ego traffic vehicles
  • Random Pedestrians: Enable for pedestrian agents
  • Random Bicyclists: Enable for bicyclists
  • Use Pre-defined Seed: Enable then input an integer to deterministically produce a random simulation that can be reproduced

Find the example guide here.

Runtime Template: Visual scenario editor#

  • Scenario: Upload the JSON file created by the Visual Scenario Editor.

Runtime Template: Python API#

  • Python Script: Input the Python API test case script that you would like to automatically execute when starting the simulation from the web user interface. Note that this is only supported for clusters running the Linux version of SVL Simulator.
  • Map: Select the map for the simulation. This must match the name of map passed to sim.load().
  • Vehicle: Select one or more ego vehicles. These must match any hardcoded vehicle names passed to sim.add_agent(). NOTE: If you have selected a cloud cluster, only a single vehicle can be selected.

Please see here for additional information on Python API mode.

Runtime Template: API Only#

For API Only simulations, you do not need to set any starting parameters, as the runner and test case you are expected to execute should determine the starting configuration. This could be from Visual Scenario Editor, or a Python API test case script.

See the below documents for additional information on running the various types of test cases using the API Only runtime template:

Autopilot top#

Local cluster top#

  • Autopilot: A list of compatible Autopilot systems are highlighted to choose from, based on the ego vehicle and the cluster type selected on the Test case pane.
  • Bridge IP: the IP address and port number that your bridge is running at, e.g. localhost:9090 or 127.0.0.1:9000.

Publish top#

Finally, click "Publish" to publish the new simulation to your library and to be able to execute it.

You can now run your simulation. Make sure the cluster you specified for the simulation is online, then click "Run Simulation".