When collaborative robots interact with people in handoff or shared-workspace tasks, motion planners frequently encounter an actuator latency penalty: by the time a human hand breaches a reactive tripwire or spatial bounding box, physical motors and joint gearboxes take 150–300 ms to spool up from a dead stop, resulting in perceptible hesitation.
Human movement exhibits an anticipatory lead time. Research in human factors and visual perception shows that ocular fixation and early reach kinematics lock onto target zones roughly 200–300 ms before physical contact.
SI Module 01 decodes non-contact optical gaze and early reach vectors at the edge using standard camera streams, broadcasting 60 Hz advisory intent directives over local WebSockets. This enables downstream ROS 2 controllers and MoveIt 2 pipelines to pre-align joints and pre-spool actuators before the human hand reaches the exchange boundary.
To allow the robotics community to evaluate the telemetry architecture and wire contract, the client integration bridge and technical specification have been released under Apache 2.0:
GitHub Repository: GitHub - Neuropatic-Labs/si-module-01-bridge: Client integration bridge, 60Hz telemetry schema, and ROS 2 adapters for SI Module 01 pre-motor intent engine. · GitHub
The open-source repository includes:
• 60 Hz WebSocket JSON telemetry schema (telemetry.v1.json)
• Synthetic 60 Hz telemetry generator for local testing (mock_telemetry_server.py)
• Reference ROS 2 subscriber and 33.3 ms watchdog bridge node (si_module_01_bridge.py)
• Technical white paper detailing the kinematic model and systems architecture (The Architecture of Intent)
While our core edge inference container (si-module-01:rc1) remains proprietary for on-premise evaluation pilots, the client bridge and synthetic server can be run locally against your existing ROS 2 workspace in a few commands:
git clone [https://github.com/Neuropatic-Labs/si-module-01-bridge.git](https://github.com/Neuropatic-Labs/si-module-01-bridge.git)
cd si-module-01-bridge
python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
python3 -m bridge.mock_telemetry_server