AI-Powered Robot based on an LBM
Use Case Family
Business Domain
R&D
Processes
Product Innovation
Challenge
Humanoid robots currently require time-intensive manual programming for each task, limiting their adaptability and scalability—especially for complex interactions involving full-body movement and object manipulation in real-world settings.
Solution
With Large Behavior Models (LBMs), a humanoid robot like Atlas learns and autonomously performs a wide range of complex tasks based on a few human demonstrations. A single neural network controls the entire body, including locomotion, balance, lifting, and manipulation.
Source: Toyota Research Institute (TRI)
Benefits
- Rapid skill acquisition without manual programming
- Whole-body control for real-world physical interactions
- Scalable development of general-purpose assistive robots
Target Group
Potential Industries
Risk Classification (EU AI Act)
High Risk systems
Art. 50; potential Art. 6 i.V.m. Annex III
with transparency obligations
