Level 0 Developer Guide
Level 0 gives you direct access to joints and sensors — position, torque, and gain control per joint, along with raw IMU and leg state data. It is the right choice if you are:
- Writing a custom locomotion controller
- Running a learned (ONNX) walking policy
- Accessing low-level sensor data directly
If you only need predefined gaits and high-level motion commands, start with the Level 1 Developer Guide instead.
Prerequisites
| Item | Details |
|---|---|
| Safety & operation | Complete the Common Guide first |
| Dev environment | Set up with the Developer Setup Guide |
| Concepts | Read Concept — focus on LV0 / Low Level |
Getting Started (Suggested Order)
- Developer Setup Guide — Clone the repo, install deps, build, and run the simulator.
- RBQ SDK — Install the SDK that provides DDS-based communication with the motion stack.
- rbq_low_level — Build and run the LV0 sample: subscribes to IMU/joints, publishes motion references, and optionally runs an ONNX walking policy.
- Verify baseline behavior in simulation, then extend only the features you need.
What You Can Do at Level 0
| Capability | Details |
|---|---|
| Per-joint control | Set position reference, torque, Kp/Kd gains via JointControl |
| Low-level state | Read IMU, joint positions/velocities/torques |
| Joystick input | Access gamepad inputs through the SDK |
| RL policy | Run ONNX inference; publish MotionRef with policy outputs |
| Motion ownership | Take/release per-joint control ownership safely |
Related Documentation
| Topic | Link |
|---|---|
| Architecture & concepts | Concept |
| API reference | API Overview |
| ROS 2 integration | ROS 2 Quick Start |
| RL training environment | RBQ GYM |
| Custom simulation | RBQ Lab |
| Safety during development | General Safety Guidelines |