Fourier Intelligence Unveils New Generation General-Purpose Humanoid Robot GR-2 with Advanced Features
TapTechNews September 26th news, Fourier Intelligence today released its independently developed new-generation general-purpose humanoid robot GR-2. The official stated that its product vision is to 'create the best embodied carrier for AI', with more flexible, more powerful, and more open features.
TapTechNews summarizes the main information of this robot as follows:
Its appearance follows the bionic design language of GR-1. After upgrading, the height reaches 175 cm (about 5.74 feet), and the weight is 63 kg (about 138.89 pounds), and there are a total of 53 degrees of freedom throughout the body. The single-arm movement load reaches 3 kg, enabling more complex operations. In addition, the capacity of the new-generation battery carried by this robot has doubled, and the endurance time has been extended to 2 hours (about 120 minutes). On the basis of supporting direct charging, it also adds a removable battery swap solution to fully meet the movement needs.
This robot is also equipped with a new-generation dexterous hand. The degree of freedom of a single hand has increased from 6 of the first generation to 12, claiming to be able to perfectly replicate the shape and size of the human hand and simultaneously simulate the natural movement of the human hand to the greatest extent.
It supports VR remote operation control, drag-and-teach, upper computer end control and other upper limb teaching methods. At the same time, it can quickly create actions and record multi-dimensional data such as upper limb movement trajectory, joint angle, force, dexterous hand movement, and tactile information, which is convenient for collecting more real training data.
According to the introduction, this robot forms a set of development interface schemes based on mainstream programming languages, supports the algorithm program development of the server-client model, encapsulates a series of simple and easy-to-use APIs, and integrates machine vision, path planning, force control feedback and other pre-optimized algorithm modules. While reducing the development threshold, it simplifies the implementation process of complex tasks and significantly improves the development efficiency. Currently, it supports NVIDIA IsaacLab, ROS, Mujoco, Webots and other open source frameworks.
In addition, according to the report of The Paper, Gu Jie, the founder and CEO of Fourier Intelligence, said in an interview, 'The GPT moment of a truly general-purpose robot has not arrived yet... But the dawn has been seen. It's not a matter of 10 or 20 years. It's within three to five years.'