Robot Motion (Experimental)

Tool Reference

This section covers the experimental Motion Generation API, the libraries and controllers that build on it, and behavior-programming workflows that coordinate them. The API provides a unified framework for real-time motion generation across wheeled robots, manipulators, humanoids, and other robots.

Section	Description
Quick Start	Get a robot moving in minutes with a ready-made controller. Covers a mobile robot path (differential drive) and a robot arm path (cuMotion RMPflow), both interactive.
Motion Generation API	The unified framework underlying all controllers and integrations. Covers RobotState control spaces, the BaseController interface, controller composition, trajectory planning, and scene interaction.
Controllers	Ready-made GPU-accelerated controllers. Currently, mobile robot controllers are implemented here (DifferentialDriveController, AckermannController, HolonomicController) built on the Motion Generation API. Future versions will include additional controllers for manipulators.
cuMotion Integration	GPU-accelerated motion generation for manipulation. Covers RMPflow reactive control, graph-based motion planning, trajectory generation, and trajectory optimization.
PINK Integration	Differential inverse kinematics via quadratic programming. Covers weighted multi-task IK, safety barriers (self-collision, workspace bounds), and QP solver selection.
State Machines for Robot Behavior	Structure reactive manipulation tasks with finite state machines, py_trees, or Omniverse Behavior Tree workflows.