Server and Client on Two Machines

By default, Carla server and client run on the same machine. However, due to the convoluted nature of the Virtual Reality environment, it is highly recommended that the server and the client being seperated on two machines. For more information on Carla server and client, check here.

In our implementation, the server machine is responsible for: hosting Unreal Engine and the Carla project, connecting the Oculus Quest headset and rendering VR animation. The client machine is responsible for: hosting the Flask server, controlling the G29 wheel, and running the PythonAPI. For more information, refer to our Running XR Simulation page.

Step 1: Connnect Two Machines

Conenct the two machines with an Ethernet cable. Due to enormous data transfer rate, CAT6 and above is required. An USB-C port with ethernet adaptor is also feasible.

Step 2: Configure IP Address

  • On the server (Unreal Engine) machine, navigate to Control Panel > Network and Internet > Network and Sharing Center. Click Ethernet > Properties > Internet Protocol Version 4 (TCP/IPv4). Manually enter the IP address 192.168.1.21, subnet mask 255.255.255.0, and leave Default gateway as blank. On the bottom, manually enter the Preferred DNS server 8.8.8.8, and leave the Alternate DNS server as blank.

  • On the client (PythonAPI) machine, navigate to Control Panel > Network and Internet > Network and Sharing Center. Click Ethernet > Properties > Internet Protocol Version 4 (TCP/IPv4). Manually enter the IP address 192.168.1.20, subnet mask 255.255.255.0, and leave Default gateway as blank. On the bottom, manually enter the Preferred DNS server 8.8.8.8, and leave the Alternate DNS server as blank.

Step 3: Configure Unreal Engine (Server)

Inside the UE4 Editor, on the top tool bar, click Edit > Project Settings. Under Plugins, find TCP Messaging. Set the Listen Endpoint to be 192.168.1.20:2000.

Inside the UE4 Editor, on the Content Browser on the bottom, navigate to Content > CarConfigurator > Car > Blueprints. Double click to open BP_AudiA5_CARLA_V3. Click EventGraph on the left panel, and drag the blueprint to find Va Rest Subsystem on the top left section. Locate the block Call URL, and change the URL from http://127.0.0.1:5000/get_steer to http://192.168.1.20:5000/get_steer.

Connect Oculus Quest 2 headset to the server machine and enable Oculus Link. Verify that the headset is successfullly conencted in the Oculus desktop application and you are able to run the simulation in the Unreal Carla project.

Step 4: Configure PythonAPI (Client)

In the PythonAPI script, find one of the following lines 

client = carla.Client(‘127.0.0.1’, 2000)

client = carla.Client(‘localhost’, 2000)

and replace it with 

client = carla.Client(‘192.168.1.21’, 2000)

This allows the client to actively look for an outside server instead of the local machine. Make sure port 2000 is not blocked by some other tasks or the windows firewall.

If a Python script involves posting data to Flask server, you need to find the following line

url = http://127.0.0.1:5000/set_steer

and replace it with

url = http://192.168.1.20:5000/set_steer

Note: You only need the PythonAPI folder on the client machine with related dependencies installed. You do not need the Carla project source code or the Unreal Engine on the client machine.

Step 5: Flask Server (Client)

Follow instructions on Flask server from the Running XR Simulation page. If the machines were connected correctly, the Flask server should simutaneously host two sites at 127.0.0.1:5000 and 192.168.1.20:5000. However, if the data request fails, manually check these sites in your browser.

Step 6: G29 Wheel (Client)

Connect G29 wheel to the client machine, and follow instructions as in the Running XR Simulation.