Boston Dynamics’ renowned four-legged robot, Spot, has received a significant intelligence boost through an advanced integration with Google’s Gemini artificial intelligence. This pivotal upgrade empowers the robot with enhanced capabilities to understand its surroundings, interpret diverse tasks, and make independent decisions with substantially less human intervention.
The robotics arm of Hyundai Motor Group recently unveiled a demonstration video on its YouTube channel, showcasing the upgraded Spot in action. The footage highlights Spot’s ability to leverage its onboard cameras and Google’s sophisticated visual language model, Gemini Robotics-ER 1.6, to accurately read and comprehend a handwritten to-do list on a chalkboard.
In the captivating video, Spot efficiently carries out various household tasks detailed on the list. These include neatly organizing scattered shoes in a rack, retrieving an empty can for disposal in a trash bin, and collecting clothes from the floor to place them in a laundry basket. Demonstrating its versatility, Spot also inspects a mousetrap hidden beneath furniture and, remarkably, takes a dog for a walk.
A separate video further illustrates Spot’s potential, depicting the robot executing more intricate supervisory and inspection tasks within a challenging manufacturing facility environment, underscoring its robust industrial applications.
Boston Dynamics confirmed that these notable performance gains are a direct result of the seamless integration of Google’s Gemini Robotics model with Boston Dynamics’ proprietary robot software program, Orbit, alongside its advanced AI Visual Inspection Learning feature.
According to Boston Dynamics, Gemini’s enhanced reasoning capabilities now enable Spot to perform a wider array of sophisticated tasks with immediate proficiency. By meticulously analyzing data gathered through Spot’s various sensors in conjunction with Gemini, the robot can achieve a deeper and more nuanced understanding of complex environments, accurately assess situational contexts, and precisely interpret the specific requirements of any given task.
sahn
