The ALOHA Unleashed method is based on ours ALOHA 2 platform based on the original ALOHA (low-cost open source hardware system for two-handed teleoperation) z Stanford University.
ALOHA 2 is much more dexterous than previous systems because it has two hands that can be easily operated remotely for training and data collection purposes, and it allows robots to learn how to perform new tasks with fewer demonstrations.
In our latest system, we have also improved the ergonomics of the robot hardware and improved the learning process. First, we collected demonstration data by remotely controlling the robot's behavior to perform difficult tasks such as tying shoelaces and hanging T-shirts. We then applied the diffusion method, predicting the robot's actions based on random noise, similar to how our Imagen model generates images. This helps the robot learn from the data so that it can perform the same tasks on its own.
Learning robot behavior from several simulated demonstrations
Controlling a dexterous robotic hand is a complex task that becomes even more complex with each additional finger, joint and sensor. In another new paperwe present DemoStart, which uses a reinforcement learning algorithm to help robots acquire skillful behaviors during simulation. These learned behaviors are particularly useful for complex embodiments such as multi-fingered hands.
DemoStart first learns from easy states and over time begins to learn from more difficult states until it masters the task as best it can. Learning how to solve a task in a simulation requires 100 times less simulated demonstrations than what is typically needed when learning from real-world examples for the same purpose.
The robot achieved a success rate of over 98% in a variety of different simulation tasks, including reorienting cubes of a specific color, tightening a nut and bolt, and organizing tools. In a real-world setup, it achieved 97% success in reorienting and lifting the cube and 64% in a plug-and-socket insertion task that required the coordination and precision of the big fingers.
