#Bird-Legged Biped Robot

A bird-legged biped robot designed with generative AI, modeled in Blender, simulated in MuJoCo, and trained using reinforcement learning to walk.

##Overview

This project focuses on building a bird-legged biped robot from design to training. The robot was first modeled in Blender, then integrated into a simulation environment, and finally trained using a reinforcement learning approach to achieve walking behavior.

##Features

• 3D robot design created in Blender
• Bird-legged biped structure
• Simulation-based training
• Reinforcement learning for walking behavior
• Visual outputs and project results included

##Files Included

• bird_biped.blend - Blender model of the robot
• train.ipynb - training notebook
• ppo_bird_biped/ - trained model files
• photos of blender/ - render images
• project report and demo outputs

##Tools and Technologies

• Blender
• MuJoCo
• Reinforcement Learning
• Stable-Baselines3
• PyTorch
• Python

##Outcome

The project demonstrates the use of generative AI, 3D modeling, simulation, and reinforcement learning to design and train a walking biped robot.