Generate captions, optimize content, and enhance with custom images for maximum engagement.
46M visits/mo-10%#3 Image & Design
Axonnic
Records human dexterous work with an EMG sleeve to train robots for automation
About Axonnic
Axonnic records how expert workers perform manual tasks using an electromyography (EMG) sleeve that captures muscle activation and inferred effort alongside video and hand pose. The sleeve leaves the hand unobstructed and measures signals upstream of contact, revealing preparation and muscle use that cameras miss. Collected data is aligned into a task-specific dataset and retargeted into the robot’s action space to train policies for delicate or force-sensitive operations. After training, the robot operates using its own sensors without the sleeve. The system targets tasks that resist automation, such as sewing deformable materials, delicate assembly, and construction work where rhythm and adaptation are critical. It is designed to capture real production scenarios rather than staged demonstrations, providing aligned EMG, video, and pose data for robotics and AI research.
Key features
- EMG sleeve for forearm signal capture
- Real-time finger tracking from muscle signals
- Alignment of EMG, video, and hand pose data
- Retargeting of human demonstrations to robot policies
- Support for delicate and deformable material handling
- Deployment without the sleeve after training
- Physics-based AI models for policy training
- Cross-domain applicability (manufacturing, construction)
Use cases
- Training robots for sewing deformable materials
- Automating delicate assembly tasks in manufacturing
- Recording construction workers for field automation
Pros
- Captures muscle activation and effort signals missed by cameras
- Leaves hands unobstructed for natural movement
- Aligns EMG, video, and pose into a single dataset
- Retargets human demonstrations into robot policies
- Targets tasks that cameras and force sensors struggle with
Cons
- Requires a waitlist to access
- Limited to tasks where human experts can wear the sleeve
- EMG data collection depends on specific hardware
Frequently asked questions about Axonnic
What does Axonnic do?
Axonnic records how expert workers perform manual tasks using an electromyography (EMG) sleeve that captures muscle activation and inferred effort alongside video and hand pose. The collected data is aligned into a task-specific dataset and retargeted to train robot policies for delicate or force-sensitive operations.
Who is Axonnic designed for?
Axonnic is designed for manufacturers and robotics labs seeking to automate dexterous tasks that resist traditional automation, such as sewing deformable materials, delicate assembly, and construction work where rhythm and adaptation are critical.
How does the EMG sleeve work?
The EMG sleeve measures muscle activation and inferred effort upstream of contact, revealing preparation and muscle use that cameras miss. It tracks individual finger movements and fingertip pressure in real time from forearm signals without obstructing the hand.
Can the robot operate without the sleeve after training?
Yes, after training, the robot operates using its own sensors without the sleeve. The EMG data guides the training process, but the robot deploys independently once the policy is learned.
What types of tasks has Axonnic been tested on?
Axonnic has been tested on sewing deformable materials, delicate assembly, machine tending, tool use, and construction work. These tasks were chosen because they are particularly challenging for traditional automation approaches.
How is the data collected and processed?
Data is collected by having expert workers perform tasks while wearing the EMG sleeve, with video and hand pose recorded simultaneously. The EMG, video, and pose data are aligned into a task-specific dataset, which is then retargeted into the robot’s action space to train policies.
