The company's CEO, Robert Playter, emphasized the importance of a robot's ability to recover from a fall, especially in real-world conditions. This has been demonstrated in the company's Spot robots, which have logged thousands of hours in real-world conditions. The company has also been working on reinforcement learning to help fallen robots right themselves. The ability for robots to recover from falls is seen as crucial for their integration into existing workflows in factories and warehouses.
Key takeaways:
- Boston Dynamics has been focusing on the ability of their robots, particularly the Atlas and Spot models, to recover from falls, acknowledging that falls are an inevitable part of operating in real-world environments.
- The company has learned a lot about falling and recovery from their quadrupedal robot, Spot, which has logged many hours in real-world conditions.
- Agility Robotics, another company in the field, has also been working on how their robot, Digit, can recover from falls, using reinforcement learning to help the robot right itself.
- Both companies emphasize the importance of robots being able to recover from falls in order to function effectively in existing workflows, particularly in factories and warehouses that were not custom-built for automation.