AI's Rapid Evolution: From Robotic Dogs to Alien-Looking Creations in Seconds
Researchers at Stanford University and Shanghai Qi Zhi Institute have developed a vision-based algorithm that enables robodogs to navigate challenging environments with agility. The algorithm combines perception and control using images from a depth camera and machine learning to move the legs and overcome obstacles. The robodogs demonstrated the ability to climb obstacles, jump over gaps, crawl under barriers, and squeeze through narrow spaces. The researchers plan to incorporate real-world data into their algorithm to enhance its autonomy.