ROAR_PY

https://roar.gitbook.io/roar_py-documentation
Admin
ROAR_PY is the ROAR team's control suite for vehicles and robots, providing an abstract layer to...
Tokens:4,241
Snippets:44
Trust Score:10
Update:9 months ago
Show doc for...
Tokens:
Raw
### Clone and Install ROAR_PY Repository

Source: https://roar.gitbook.io/roar_py-documentation/interface-installation

This sequence of commands clones the ROAR_PY repository from GitHub and then installs the core ROAR_PY package and its dependencies using pip. It requires navigating into the cloned directory first.

```bash
git clone https://github.com/augcog/ROAR_PY
cd ROAR_PY/roar_py_core
pip install -e .
```

--------------------------------

### Verify Python Version in Conda Environment

Source: https://roar.gitbook.io/roar_py-documentation/interface-installation

After activating the conda environment, this command verifies that Python 3.8 has been successfully installed and is the active version within the environment.

```shell
python --version
```

--------------------------------

### Setup Asynchronous Mode in ROAR Py

Source: https://roar.gitbook.io/roar_py-documentation/miscellaneous/initial-design-document

This snippet demonstrates how to set up the CARLA client and ROAR Py server in asynchronous mode. It configures the world to not wait for client actions before ticking and establishes a WebSocket server to handle client connections and messages.

```python
connected_clients = []

def action_handler(ws,action):
    world.apply_action(action)
    if not_throttled:
        observation = world.get_observation(ws)
        ws.send(observation)
    elif too_much_throttled:
        kick_client(ws)


def ws_handler(ws,init_msg):
    connected_clients.append(ws)
    world.initialize_agent(init_msg)
    ws.subscribe(action_handler)

def main(_):
    client = carla.Client("localhost", 2000)
    world = client.get_world()
    world.set_synchronous_mode(False) #Set to asynchronous mode
    wss = WebSocketServer(handler = ws_handler)
    # ...
    await wss.spin() #Continue to listen to incoming connections and messages
```

--------------------------------

### Create and Activate Conda Environment for ROAR_PY

Source: https://roar.gitbook.io/roar_py-documentation/interface-installation

This snippet demonstrates how to create a new conda virtual environment named 'roar_py' with Python 3.8 installed, and then activate it. This ensures a clean and isolated environment for ROAR_PY.

```shell
conda create -n roar_py python=3.8
conda activate roar_py
```

--------------------------------

### Attach and Print Accelerometer Data in CARLA

Source: https://roar.gitbook.io/roar_py-documentation/basics/sensors

Demonstrates how to create and attach an accelerometer sensor to a vehicle in CARLA, then continuously receive and print its acceleration data. It requires the `roar_py_interface` and assumes a CARLA world setup. The output is acceleration data in meters.

```python
#create an object and attach
accelerometer = vehicle.attach_accelerometer_sensor(...) # Parameter should include location, rotation of the sensor, etc.

#print accel data
try:
    while True:
        await carla_world.step()
        accel: roar_py_interface.RoarPyAccelerometerSensorData = await accelerometer.receive_observation()
        print(accel)
finally:
    roar_py_instance.close()
    

```

--------------------------------

### Synchronous Mode Game Loop in Python

Source: https://roar.gitbook.io/roar_py-documentation/miscellaneous/initial-design-document

This Python pseudocode illustrates the game loop for the ROAR-MA-SIM-SERVER in synchronous mode. It handles client connections, retrieves observations from the CARLA world, sends them to connected clients, receives actions from clients, applies actions, and ticks the CARLA world. Dependencies include the CARLA client library and a WebSocket server implementation. It assumes a `world` object with methods like `initialize_agent`, `get_observation`, `apply_action`, and `tick`.

```python
connected_clients = []

def ws_handler(ws,init_msg):
    connected_clients.append(ws)
    world.initialize_agent(init_msg)

def main(_):
    client = carla.Client("localhost", 2000)
    world = client.get_world()
    world.set_synchronous_mode(True) #Set to synchronous mode
    wss = WebSocketServer(handler = ws_handler)
    # ...
    while True:
        # ...
        actions = []
        for client in connected_clients:
            observations = world.get_observation()
            await client.send(observations)
        for client in connected_clients:
            next_action = await client.recieve_next_action(timeout = 0.2)
            if next_action is not None:
                actions.append(next_action)
            else:
                kick_client(client)
        for action in actions:
            world.apply_action(action)
        world.tick()
```
Tokens:
Raw