A multi-sensory robot brain
NEURA Robotics has introduced Neuron AI, a foundation model designed to help robots understand their surroundings and carry out physical tasks. The company describes it as a Perception-Language-Action architecture that turns sensory input into action. The system combines 3D vision, tactile and force measurements, audio, robot-state data, language and memory in a single model.
Why touch and sound matter
According to NEURA, tactile and force data can help a robot judge whether a physical interaction produced the intended result. For example, a change in contact force could indicate whether a screw has engaged correctly, while audio cues could reveal that a mechanical latch has clicked into place. The company presents these as intended capabilities rather than independently verified performance results.
Five embodiments, on-device inference
The model supports five robot embodiments, including the 4NE1 Gen 3 and Gen 4 humanoids, a wheeled 4NE1 platform and two additional bimanual platforms. NEURA says inference runs locally on the robot without a cloud connection. Training combines real-world robot experience, simulated variations and human demonstrations recorded with sensor-equipped suits.
Ecosystem and demo plans
The company collects data, trains skills and benchmarks them through its NEURA Gyms infrastructure, and cites industrial data partnerships with Schaeffler and Bosch. Through the Neuraverse platform, customers can fine-tune robot capabilities for their own workflows and distribute them across fleets. NEURA plans to demonstrate Neuron AI with its 4NE1 humanoid at NVIDIA GTC Berlin from October 20-22.
