SCARA Pick-and-Place with Reinforcement Learning

This project simulates a SCARA robot performing a pick-and-place task using PyBullet and trains it using PPO from Stable-Baselines3.

🚀 Features

• SCARA robot simulated in PyBullet
• Cube pick and place using inverse kinematics
• Reward-based PPO training (Stable-Baselines3)
• Replay, evaluation, and success metrics
• Headless and GUI-compatible

🧰 Requirements

pip install -r requirements.txt

🏁 How to Train

python scara_train_checkpointed.py

🎬 How to Evaluate Pick-and-Place

python scara_pick_place_eval.py

📁 Project Structure

scara-pickplace-rl/
├── scara_orange.urdf
├── scara_pick_env_final.py
├── scara_pick_env_place.py
├── scara_train_checkpointed.py
├── scara_pick_place_eval.py
└── scara_rl_logs/
    └── scara_model_step_*.zip

📜 License

MIT License