3 tracks Β· 18 modules Β· From concept to deployment
What VLA is: models, datasets, hardware, safety
How-to: Python, LeRobot, simulators, ROS2
Where itβs headed: dual-system, world models, market, roadmap
From LLMs to Robots That Act: The Big Picture
Anatomy of a VLA: Vision + Language + Action
Model Overview: RT-2, OpenVLA, Ο0, Qwen-VLA
Datasets: Open X-Embodiment and DROID
Hardware: Cobots, Humanoids, and Platforms
Safety and Standards: ISO 10218:2025
Python ecosystem: PyTorch, Transformers, Gymnasium
LeRobot: End-to-End Pipeline
Simulators: MuJoCo, Isaac Sim, PyBullet
ROS2: Middleware for Real Robotics
Training: Imitation Learning, ACT, Diffusion Policy
Sim-to-Real: From Virtual to Physical
Dual-System Architecture (S1/S2)
World Models and Predictive Simulation
Cross-Embodiment Generalization
Sector Applications: Logistics, Healthcare, Agriculture
Market and Opportunities: $7.2B
Roadmap 2027-2030: The Future of Embodied AI