Ambi Robotics Leverages Sim2Real Technology to Power Picking Robots

Analysis:
Pick-and-place robots for logistics applications are not novel, but Ambi’s operating system is. The robots are powered by AmbiOS, the company’s advanced operating system that leverages proprietary simulation-to-reality technology to rapidly teach robots unique pick and pack tasks. AmbiOS leverages The Dexterity Network (Dex-Net) project that automates the training of deep neural networks to improve a robot’s ability to grasp items. Dex-Net was created by Goldberg and his team at the University of California, Berkeley. AmbiOS trains robots in simulation and transfers its deep neural network advancements to real-world systems. Goldberg claims Dex-Net trains robots in hours, not months, 10,000 times faster than collecting data in the real world. Faster training in simulation opens up new opportunities for robotic manipulation, creating more capable robots from day one that can help companies handle surging e-commerce demand. – Steve Crowe

---