---
license: apache-2.0
library_name: transformers
pipeline_tag: robotics
tags:
  - reinforcement-learning
  - robotics
  - physical-ai
  - mujoco
  - ppo
  - bird
  - robot-bird
  - bipedal-vtol
  - lerobot
  - autonomous-agents
code: https://github.com/artnfull-bot/OpenBird-artnfull
model-index:
  - name: open-bird-robot-mujoco-ppo
    results:
      - task:
          type: reinforcement-learning
          name: Avian Bipedal Walking & 180-deg VTOL Flight
        dataset:
          name: OpenBird 15-Stage Multi-Modal Trajectory Dataset v1.0
          type: robotics-demonstration
        metrics:
          - type: mean_reward
            value: 120.0
            name: 15-Stage Full Sequence Completion
          - type: success_rate
            value: 100.0
            name: VTOL Takeoff & Landing Rate (%)
---

# 🦅 OpenBird (open-bird-robot-mujoco-ppo) AI Models & Motion Datasets

[![Hugging Face Model](https://img.shields.io/badge/🤗%20Hugging%20Face-Model%20Hub-orange.svg)](https://huggingface.co/artnfull/open-bird-robot-mujoco-ppo)
[![Hugging Face Space](https://img.shields.io/badge/🤗%20Hugging%20Face-Spaces-yellow.svg)](https://huggingface.co/spaces/artnfull/OpenBird-artnfull)
[![GitHub Repository](https://img.shields.io/badge/GitHub-OpenBird--artnfull-blue?logo=github)](https://github.com/artnfull-bot/OpenBird-artnfull)
[![License: Apache 2.0](https://img.shields.io/badge/License-Apache_2.0-blue.svg)](https://opensource.org/licenses/Apache-2.0)
[![Design License: CC BY-NC-SA 4.0](https://img.shields.io/badge/Design-CC_BY--NC--SA_4.0-lightgrey.svg)](https://creativecommons.org/licenses/by-nc-sa/4.0/)

> **Official Physical AI simulation platform, pre-trained policy weights, and 15-stage trajectory datasets for the OpenBird-artnfull Bipedal VTOL Robot Platform.**  
> 🎮 **Playable Interactive Simulator**: Download from GitHub or Hugging Face and directly control the robot bird via keyboard (WASD)!

---

## 🌟 Key Features & Kinematics

- 🦿 **Digitigrade Bipedal Locomotion**: Avian joint kinematics with dynamic reverse-knee walking and counter-balancing.
- 🪽 **180° Deployable Duct Wings**: Stowed vertically behind the back during walking; deployed horizontally (180°) for high-thrust VTOL flight.
- 🦅 **Look-Down Eye-Gaze & Beak Pitch Control**: Independent 60° downward ground scan without disturbing body equilibrium.
- 📷 **Multi-Modal Vision Perception**: Stereo Eye Cameras + Cockpit FPV view.

---

## 🎮 Interactive Keyboard Control Guide

You can download and directly control the robot bird locally in real-time 3D simulation!

| Key | Action | Description |
| :---: | :---: | :--- |
| `A` / `D` | **Yaw Turn** | Turn left / right (Walking turn / Flight yaw) |
| `↑` / `↓` (`W`/`S`) | **Forward / Backward** | Walk forward/backward (Ground) or Fly forward/backward (Air) |
| `←` / `→` | **Strafe** | Sideways translation in air |
| `R` / `F` | **Altitude** | Altitude Climb (+15cm) / Descend (-15cm) |
| `SPACE` | **Stop / Hover** | Stand still (Ground) / Precision Hover (Air) |
| `1` $\rightarrow$ `2` $\rightarrow$ `3` | **Flight Sequence** | Deploy Wings $\rightarrow$ VTOL Takeoff & Hover $\rightarrow$ Soft Landing & Fold |
| `C` | **Camera Switch** | 3rd-Person Orbit $\rightarrow$ Head FPV $\rightarrow$ Left Eye $\rightarrow$ Right Eye |
| `V` / `B` | **Eye-Gaze Pitch** | Look down 60° (`V`) / Look forward (`B`) |
| `G` / `Y` / `K` / `P` | **Gestures** | Peck (`G`), Crouch (`Y`), Knockdown (`K`), Auto-Right (`P`) |
| `Q` / `ESC` | **Exit** | Close simulator |

---

## 🚀 How to Download & Run Locally

### Option A: From GitHub (Recommended)
```bash
# 1. Clone repository
git clone https://github.com/artnfull-bot/OpenBird-artnfull.git
cd OpenBird-artnfull

# 2. Install dependencies (MuJoCo, NumPy, OpenCV)
pip install -r requirements.txt

# 3. Launch Interactive Flight (Manual Control)
run_manual.bat   # or python run_robot_bird.py

# 4. Launch Autonomous 15-Stage Cinematic AI
run_ai_auto.bat  # or python run_ai_robot_bird.py
```

### Option B: Download via Hugging Face Hub (Python)
```python
from huggingface_hub import snapshot_download

# Download complete model assets, datasets, and simulator scripts
repo_path = snapshot_download(repo_id="artnfull/open-bird-robot-mujoco-ppo", repo_type="model")
print(f"Downloaded OpenBird assets to: {repo_path}")
```

---

## 📂 Repository Contents

```text
OpenBird-artnfull/
├── policies/               # Trained neural network weights (.pt, .onnx)
│   └── baseline_flight/    # Baseline VTOL takeoff, hover, and landing models
├── datasets/               # MuJoCo / LeRobot compatible motion trajectories
│   └── demonstration/      # Teleoperation & autonomous behavioral data
├── outputs/eval/           # Official evaluation metrics & trajectory logs
│   ├── eval_info.json      # Rollout evaluation statistics
│   └── videos/             # High-definition evaluation rollouts
├── robot_bird/             # MuJoCo 3D kinematics, FSM controller & XML models
├── run_robot_bird.py       # Interactive Keyboard Flight simulator
├── run_ai_robot_bird.py    # 15-Stage Autonomous Cinematic Showcase
├── run_manual.bat          # Windows 1-Click Interactive launcher
├── run_ai_auto.bat         # Windows 1-Click Autonomous launcher
├── requirements.txt        # Dependency list (mujoco, numpy, opencv-python)
└── README.md               # Model & dataset card
```