Boston Dynamics has announced that its quadruped robot, Spot, is now equipped with Google DeepMind’s Gemini Robotics-ER 1.6, a high-level embodied reasoning model designed to enhance the robot's usability and intelligence for complex tasks. This development, revealed today, marks a significant advancement in the commercial deployment of legged robots, particularly in industrial inspections, where Spot will autonomously identify hazardous debris, read gauges, and utilize vision-language-action models for better environmental understanding.
The collaboration aims to improve how robots interpret and interact with their surroundings, addressing the challenges of ensuring that robotic actions align with human reasoning. Marco da Silva, vice president of Spot at Boston Dynamics, emphasized that the new capabilities will allow Spot to autonomously navigate real-world challenges more effectively.
Despite the progress, experts acknowledge ongoing challenges in achieving seamless human-robot interaction. Carolina Parada from Google DeepMind noted that while the Gemini model enhances visual recognition, it currently lacks integration with other sensory data, such as touch, which is crucial for reliable object manipulation. As part of the deployment, customers using Spot for inspections will need to share operational data with Boston Dynamics to further refine the technology.
The introduction of Gemini Robotics-ER 1.6 is seen as a step toward creating safer and more reliable robots capable of performing everyday tasks, with the potential to apply these advancements to other robotic platforms in the future.
Leave a comment