UC Berkley researchers unveil the Berkley Humanoid, a low-cost, mid-scale humanoid robot for learning-based control designed specially to learn algorithms with low simulation complexity. It’s capable of anthropomorphic motion and can withstand heavy kicks without falling over.
Its narrow sim-to-real gap allows for agile and robust locomotion across various terrains in outdoor environments, thanks to a simple reinforcement learning controller using light domain randomization. Berkley Humanoid can traverse for hundreds of meters, walking on a steep unpaved trail, and even hop around with single and double legs.
- Keep track of and find your items alongside friends and devices in the Find My app
- Key finder and locator for your wallet, luggage, backpack, and more. Share an AirTag with up to 5 people, so items that everyone uses can be tracked...
- Simple one-tap setup instantly connects AirTag with your iPhone or iPad. System require: iPhone SE, iPhone 6s, or iPod touch (7th generation) with iOS...
Capable of omnidirectional locomotion and withstanding large perturbations with a compact setup, our system aims for scalable, sim-to-real deployment of learning-based humanoid systems,” said the researchers.
