Motion Generation

Lula is a high-performance motion generation library for robotic manipulation. RMPflow provides real-time, reactive local policies to guide a robot manipulator to a task space target while avoiding dynamic obstacles. A suite of Rapidly-exploring Random Tree (RRT) algorithms, including RRT-Connect and JT-RRT, deliver global planning solutions in static environments. Additionally, the trajectory generation tools in Lula provide time-optimal trajectories for paths described as a series of c-space and task-space moves. Finally, Lula provides interfaces to the performant forward and inverse kinematic solvers underpinning the higher-level motion generation tools.

Omniverse Isaac Sim also interfaces with cuRobo, a high-performance, GPU-accelerated robotics motion generation library that adds additional features to Omniverse Isaac Sim such as batched collision-free inverse kinematics, collision-free motion planning, and reactive control in the presence of obstacles represented as meshes or nvblox maps. For more information, see the cuRobo tutorial.

• Lula Robot Description and XRDF Editor
• Lula RMPflow
• Lula RRT
• Lula Kinematics Solver
• Lula Trajectory Generator
• Configuring RMPflow for a New Manipulator
• OmniGraph: Pick-and-Place Controller Node
• cuRobo and cuMotion

Examples

• Follow Target Standalone Example: standalone_examples/api/omni.isaac.franka/follow_target_with_rmpflow.py

• Follow Target Standalone Example: standalone_examples/api/omni.isaac.franka/follow_target_with_ik.py

• Follow Target Example: Isaac Examples > Manipulation > Follow Target

• Follow Target Example: Isaac Examples > Multi-Robot > RoboFactory

• Follow Target Example: Isaac Examples > Multi-Robot > RoboParty