🎮 Robot Control

OrcaGym provides a rich set of robot control interfaces, from low-level joint control to high-level inverse kinematics.

For complete runnable code, see OrcaPlayground examples/euler/ (Lessons 4-9).

Control Hierarchy

High-Level API
  ├── Damped Least Squares IK      ── end-effector pose → joint positions (with limit clamping)
  ├── PD Controller                ── target position → control torque
  ├── ONNX Walking Policy          ── observation → target joint angles (G1 walking)
  └── Mocap Control                ── directly set mocap pose + WELD constraint driving

Low-Level API
  ├── do_simulation()              ── set ctrl + step + auto-sync (recommended)
  ├── set_joint_qpos()             ── set joint positions
  ├── apply_body_force()           ── apply external force
  └── mj_forward()                 ── refresh derived quantities

Chapter Navigation

• 📡 State Query API — Complete API for joints, bodies, sensors, and contact queries
• 🎯 Action Space — Action space definition and types
• 👁️ Observation Space — Observation construction and normalization
• 🦿 Joint Control — set_ctrl, PD controller, low-pass filtering
• 🦾 Inverse Kinematics — Damped least squares IK + joint limits
• 🎭 Mocap Control — Mocap body + WELD constraint = object dragging
• 🔄 External Force Application and IK — External force application + Jacobian + IK complete workflow