PupperAI

Overview

This repository use robot model based on Stanford Pupper quadruped.

Robot goal is to walk as far as it can on given time.

Key Features :

• Deep Reinforcement Learning (DRL) with PPO agent.
• Transfer learning.
• Domain Randomization.

How it works

PyBullet Model

Real Robot

Evaluation

Training Performance

Testing Performance

Deploy