Skip to content

Latest commit

 

History

History
150 lines (101 loc) · 7.43 KB

README.md

File metadata and controls

150 lines (101 loc) · 7.43 KB
Raw

CARLA Simulator Ubuntu 22.04 + NICE DCV + GPU

This AWS CloudFormation template will deploy CARLA Simulator into an accelerated computing instance running the NICE DCV server.

CARLA Animation

CARLA

CARLA is an open-source simulator for autonomous driving research to support the development, training, and validation of autonomous driving systems.

NICE DCV

Architecture

NICE DCV is a high-performance remote display protocol that provides customers with a secure way to deliver remote desktops and application streaming from any cloud or data center to any device over varying network conditions. For security reasons we do not put the EC2 instance in a public subnet, therefore you need to use the AWS systems manager with the Session Manager plugin to tunnel the DCV port to your local machine.

aws ssm start-session --target i-<instance id> --document-name AWS-StartPortForwardingSession --parameters '{"portNumber":["8443"], "localPortNumber":["8443"]}'

Then setup the NICE DCV client to connect to localhost:8443 and set Connection Setting to be WebSockets/TCP instead of QUIC.

To use the simplified architecture without Transit Gateway (using template-unicast.yml), follow these steps:

  1. Be aware that this setup does not support UDP multicasting.
  2. Configure both the device simulator and CARLA for unicast when using ROS2.
  3. Find the example configuration for cyclonedds at: carla-client/ros2/cyclonedds_unicast.yaml.
  4. Before running the ROS2 example, set the CYCLONEDDS_URI environment variable with this command: 
    export CYCLONEDDS_URI=carla-client/ros2/cyclonedds_unicast.yaml

Deployment instructions

Download desired template file and login into the AWS CloudFormation console. Choose Create Stack and Upload the template file "template.yml".

Specify a Stack name and specify parameter values. All fields are required.

  • imageId : System Manager Parameter path to AMI ID.
  • instanceType: appropriate instance type.
  • ec2Name: Name of the EC2 instance.
  • BigaAmiId: AMI of the EC2 image for the Biga device. This can be found in the end of the CodeBuild build logs when the EC2AMIBigaPipeline-EmbeddedLinuxPipeline* is finished. Search for "AMI ID:".
  • BigaInstanceType: Instance type of the Biga device.
  • BigaInstanceName: Name of the Biga device.

CloudFormation parameters

Continue Next with Configure stack options, Review settings, and click Create Stack to launch your stack.

It may take up to 60 minutes to provision the EC2 instance. After your stack has been successfully created, its status changes to CREATE_COMPLETE. Go to the Outputs tab.

CloudFormation outputs

Go to the SSMSessionManager row, open the URL (in the form https://<REGION>.console.aws.amazon.com/systems-manager/session-manager/<InstanceID>) in a new browser tab to log in via SSM Session Manager to change login user password. The password change command is " sudo passwd ubuntu".

Do start the

aws ssm start-session --target <instance id> --document-name AWS-StartPortForwardingSession --parameters '{"portNumber":["8443"], "localPortNumber":["8443"]}'

Use the local DCVConsole and log in as an ubuntu user with the recently changed password. Do not forget setting to be WebSockets/TCP instead of QUIC.

Installing CARLA Simulator

Once logged in to the Ubuntu remote desktop, run the command "./install-carla" from your user's root directory. The installation will download everything needed to set up CARLA, change the default Python version to 3.8, and creates a virtual environment installing all package required. It may take 5 to 10 minutes to finish based on the large size of the downloaded binaries.

If you're facing timeouts with the Debian package server when installing CARLA, you can try this alternative method for installation:

1. Install Required System Dependency

Before downloading CARLA, install the necessary system dependency:

sudo apt-get -y install libomp5

2. Download the CARLA 0.9.13 Release

Download the CARLA_0.9.13.tar.gz file (approximately 16GB) from the official release:

wget https://carla-releases.s3.us-east-005.backblazeb2.com/Linux/CARLA_0.9.13.tar.gz

3. Unpack CARLA to the Desired Directory

Unpack the downloaded file to /opt/carla-simulator/:

tar -xzvf CARLA_0.9.13.tar.gz -C /opt/carla-simulator/

4. Install the CARLA Python Module

Finally, install the CARLA Python module and necessary dependencies:

python -m pip install carla==0.9.13
python -m pip install -r /opt/carla-simulator/PythonAPI/examples/requirements.txt

Then, you can test your CARLA installation by running "./CarlaUE4.sh" from "/opt/carla-simulator". You will see an image with the world simulation running successfully.

CARLA Running

Installing ROS

Follow instructions here.

Installing socket-can-setup

Follow instructions here.

Testing the PythonAPI

The best way to test the Python API is to drive a car. In order to do so, you will have to follow the steps below:

  • Start the server "/opt/carla-simulator/CarlaUE4.sh -no-rendering -quality-level=Epic -prefernvidia" in different terminal
  1. Activate the virtual environment created by the installation with the command ". venv/bin/activate" from the "ubuntu" user's home.
  2. Navigate to the PythonAPI examples folder "cd /opt/carla-simulator/PythonAPI/examples".
  3. Run "python manual_control.py".

Here is the link to the complete CARLA documentation

CARLA PythonAPI

This AWS CloudFormation template has been made possible by using as a reference the Amazon EC2 NICE DVC Samples.

Testing vcan connection to biga ec2 device

Running this on the biga ec2 instance, connect via EC2 serial console. Stop after test - "ctrl-c".

candump vcan0

Running this on the Carla ec2 instance, in a terminal.

./multicast_can_send.sh

Running this on the Carla ec2 instance, in another terminal.

cansend vcan0 123#00FFAA5501020304

You should see this in the biga ec2 instance:

root@ip-10-0-1-93:~# candump vcan0
vcan0  123   [8]  00 FF AA 55 01 02 03 04

Security

See CONTRIBUTING for more information.

License

This library is licensed under the MIT-0 License. See the LICENSE file.