How to Use Sensor

This tutorial guides you on how to use a sensor to retrieve corresponding observation data.

Pre-defined Sensors

The directory grutopia_extension/sensors/ contains a list of all available sensors::

grutopia_extension/
└── sensors
    ├── camera.py
    ├── mocap_controlled_camera.py
    └── rep_camera.py
    ...

For each robot, we provide some ready-to-use sensor configurations for each robot in grutopia_extension/configs/robots/{robot_name}.py.

How to Use a Sensor

Typically, a sensor should be used with a robot. So first of all, the sensor configuration should be added to the sensor list in robot configuration:

from grutopia.core.config import Config, SimConfig
from grutopia.core.gym_env import Env
from grutopia.core.runtime import SimulatorRuntime
from grutopia.core.util import has_display
from grutopia.macros import gm
from grutopia_extension import import_extensions
from grutopia_extension.configs.robots.jetbot import (
    JetbotRobotCfg,
    move_by_speed_cfg,
    camera_cfg,
)
from grutopia_extension.configs.tasks import (
    SingleInferenceEpisodeCfg,
    SingleInferenceTaskCfg,
)

headless = not has_display()

config = Config(
    simulator=SimConfig(physics_dt=1 / 240, rendering_dt=1 / 240, use_fabric=False),
    task_config=SingleInferenceTaskCfg(
        episodes=[
            SingleInferenceEpisodeCfg(
                scene_asset_path=gm.ASSET_PATH + '/scenes/empty.usd',
                robots=[
                    JetbotRobotCfg(
                        position=(0.0, 0.0, 0.0),
                        scale=(5.0, 5.0, 5.0),
                        controllers=[move_by_speed_cfg],
                        sensors=[camera_cfg],
                    )
                ],
            ),
        ],
    ),
)

sim_runtime = SimulatorRuntime(config_class=config, headless=headless, native=headless)

import_extensions()
# import custom extensions here.

env = Env(sim_runtime)
obs, _ = env.reset()

i = 0

while env.simulation_app.is_running():
    i += 1
    action = {move_by_speed_cfg.name: [0.5, 0.5]}
    obs, _, terminated, _, _ = env.step(action=action)

    if i % 1000 == 0:
        print(i)
        for k, v in obs['sensors'][camera_cfg.name].items():
            print(f'key: {k}, value: {v}')

env.simulation_app.close()

In the above example, first we import the move_by_speed_cfg for jetbot. It’a a ready-to-use sensor config for jetbot to use the Camera to get observations:

grutopia_extension/configs/robots/jetbot.py

camera_cfg = CameraCfg(
    name='camera',
    prim_path='chassis/rgb_camera/jetbot_camera',
    resolution=(640, 360),
)

The sensor config is then added to the robot config to declare it as an available sensor for the robot in that episode. In each step, we can read the observations from the obs dict returned by env.step(). Observations from certain sensor are stored in obs['sensors'][{sensor_name}]. The data structure of observation is defined by the get_data method of the specific sensor. For the above example, we can check it in the Camera class:

grutopia_extension/sensors/camera.py

class Camera(BaseSensor):
    def get_data(self) -> Dict:
        if self.config.enable:
            rgba = self._camera.get_rgba()
            frame = self._camera.get_current_frame()
            return {'rgba': rgba, 'frame': frame}
        return {}

So the rgba and frame data would be printed every 1000 steps in our example.