Overview

​

Key Features

• Unified Interface: Same API for real and simulated robots
• Form Factor Support:
  • Wheeled robots
  • Robotic arms
  • Legged robots
  • Aerial robots
• Environment Agnostic: Code works in both simulation and real-world
• Sensor Integration: Standardized access to cameras, LiDAR, and other sensors

​

Same Code, Different Robots

# Wheeled robot in simulation
from grid.robot.wheeled import AirGenCar

car = AirGenCar()
car.getPosition()  # Returns Position(x, y, z)
car.getOrientation()  # Returns Orientation(x, y, z, w)

# Aerial robot in real world
from grid.robot.aerial import ModalaiDrone

drone = ModalaiDrone()
drone.getPosition()  # Same method returns Position
drone.getOrientation()  # Same method returns Orientation

​

Sensor Access

# Get camera image from real robot
from grid.robot.wheeled.jetbot import JetBot

robot = JetBot()
rgb_image = robot.getImage("front_center", "rgb")

# Get camera image from simulated robot
from grid.robot.wheeled import AirGenCar

car = AirGenCar()
rgb_image = car.getImage("front_center", "rgb")

​

Common Base Classes

• Robot: Base class for all robots
  • Wheeled: For wheeled platforms
  • Arm: For robotic arms
  • Aerial: For aerial vehicles
  • Locomotion: For legged robots

​

Key Concepts

1. Common Types and APIs
2. Movement Control
   • High-level movement commands
   • Low-level control when needed
3. Sensor Integration
   • Standard image types: rgb, depth, segmentation
   • Common data formats across platforms

​

Available Robots

​

Wheeled Robots

• AirGenCar (Simulation)
• JetBot (Real)

​

Robotic Arms

• UR5e (Real)
• IsaacArm (Simulation)

​

Aerial Robots

• ModalAIDrone (Real)
• AirGenDrone (Simulation)

​

Legged Robots

• Unitree Go2 (Real)
• IsaacLocomotion (Simulation)

​

Humanoid Robots

• IsaacHumanoid (Simulation)