Today, we present a step-change in robotic AI @sundayrobotics.
Introducing ACT-1: A frontier robot foundation model trained on zero robot data.
- Ultra long-horizon tasks
- Zero-shot generalization
- Advanced dexterity
🧵->
What is the weakness of traditional model-based trajectory planners for UAVs? No flexible agility adaptation to the robot's self-awareness and observations.
Do not trust end-to-end policies? Hierarchically combining RL with a trajectory planner is a simple but effective way.