From home to work; Gemini Robotics smart robots come to the aid of humans

Google Dipmind has taken a big step towards developing useful and intelligent robots by introducing two new models of artificial intelligence based on Gemini 2.0. These models, with the names of Gemini Robotics and Gemini Robotics-ER, offer new abilities in the field of robotics and interaction with the physical world.

Models for direct control of robots

Gemini Robotics is an advanced visual-language-operation (VLA) model based on Gemini 2.0 and added physical measures as a new output mode. This model allows robots to perform more complex tasks in the real world. Gemini Robotics, using advanced Gemini language capabilities, can understand interactive commands and respond quickly to environmental changes.

Improvement of spatial understanding and visualized reasoning

The second model, the Gemini Robotics-ER, is a visual-language model that enhances the capabilities of Gemini’s visualized reasoning. This model helps robots to better understand their surroundings and perform more sophisticated tasks such as motion planning and code production. The Gemini Robotics-ER can also use intravenous learning to improve its performance.

Collaborate with leading companies in the field of robotics

Google Dipmind is working with companies like Appronik to develop the next generation of human robots. Also, these models are available to reputable companies such as Agile Robots, Boston Dynamics and Enchanted Tools to test their capabilities in real environments.

Safety and responsibility challenges

Given the importance of safety in robotics, Google Dipmand uses a multilayer approach to ensure robots safety. This approach involves low -level safety controls such as preventing collision and restricting contact forces, to understand the semantic high level for safe decisions. Also, a new dataset called ASIMOV has been released that helps researchers evaluate the safety of robots’ actions in real scenarios.

The future of smart robots

Google Dipmand hopes to develop these models to provide the ground for the emergence of a new generation of useful and intelligent robots. These robots can help humans in different environments, from home to work and perform complex tasks with high safety and safety.

With the introduction of Gemini Robotics and Gemini Robotics-ER, Google Dipmind has shown that artificial intelligence can go beyond the digital world and enter the physical world. These advances not only enhance the ability of robots, but also raise new challenges in the field of safety and responsibility that require serious attention.