Scientists Unveil a Robotic Dog Controlled Entirely by Human Thought

A robotic dog controlled purely by thought has been developed by researchers at Xi'an Jiaotong University. The team, led by Professor Xui Guanghua, combined brain-signal technology with autonomous navigation to create a system that responds to human intentions without joysticks or buttons. The robotic dog operates through a non-invasive brain-computer interface (BCI) that reads brain signals. When a user thinks of a command, the system detects the corresponding brain activity and translates it into movement. Tests show the technology achieves over 95% accuracy, with a delay of about one second between thought and action.

The dog understands 11 basic commands, such as moving forward, turning, and stopping. To improve efficiency, the team split tasks between human and machine—users focus on high-level decisions while the robot handles navigation and stability. Professor Xui's goal extends beyond simple control. His team aims to merge brain interfaces with AI and environmental sensors, creating robots that adapt to real-world conditions. The technology could eventually assist in rehabilitation, elder care, and mobility support for those with physical limitations. Advances in non-invasive BCIs have accelerated in recent years. China approved its first commercial BCI device, NEO by Neuracle Medical Technology, in March 2026 for treating paralysis. While earlier EEG headsets were basic, modern systems now use multi-channel sensors for clearer signal detection.

The mind-controlled robotic dog demonstrates how brain-computer interfaces can move beyond medical uses into everyday applications. With high accuracy and minimal delay, the system offers a practical way to control machines through thought alone. Future developments may expand its role in assistive technology and beyond.