NEURA introduces a robot model combining senses and memory
NEURA Robotics introduced Neuron AI, a model that combines visual, tactile, audio and movement information with language and memory. The company describes local operation on robots and a shared training approach across different bodies. A planned Berlin laundry demonstration will present the system’s task sequence later in October.
Artificial Intelligence··Evening
A model combines sensory information
German robotics company NEURA Robotics introduced Neuron AI on 14 October, bringing multiple sensory inputs into robot control.[1], [2]
The model combines vision, touch, audio and movement information. NEURA developed and trained the system at its facilities in Metzingen, Germany, describing it as a foundation for robots performing physical tasks.[1]
NEURA says memory preserves relevant context when an object leaves view and retains information about previous actions. Language and the robot’s own state are also inputs to the architecture.[1]
Training connects robots, simulation and people
Training draws on robot interactions, simulated environments and demonstrations by people wearing sensor-equipped suits. NEURA says the shared model serves 5 robot bodies with different hands and grippers. It describes inference running on the robot itself, allowing operation without a cloud connection. Customers can adapt skills to their tasks and distribute updates across a fleet through Neuraverse, the company’s deployment system; sharing customer skills is optional.[1]
Berlin demonstration will follow a laundry task
NEURA plans a laundry demonstration at NVIDIA’s Berlin event on 20–22 October. Its humanoid will sort clothes, load a washing machine, close and reopen it, unload the laundry and move it to a dryer. The sequence is scheduled for a future demonstration, with the announced model’s capabilities described by the company.[1]
