The Three layer system architecture
Layer 1 — Mechanics
The physical substrate that interacts with the real world.
This layer defines the system’s force limits, contact geometry, failure modes, and safe operating envelope. A stable architecture begins with hardware designed for predictable behavior under real-world physics.
Layer 2 — Peripheral Sensing
The sensing layer detects what is physically happening at the boundary between machine and environment.
Rather than forwarding raw sensor data upstream, this layer converts signals into simple physical states such as:
- stable contact
- slip
- overload
- jam
- degraded sensing
This creates a deterministic picture of physical reality in real time.
Layer 3 — Reflex Kernel
The reflex kernel is the final authority before actuation.
Every movement command passes through this deterministic enforcement layer before reaching the hardware. If a command would cause unsafe or unstable behavior, the reflex layer can:
- limit
- hold
- reject
- back off
This ensures that unsafe commands never reach the actuators.
Why the Layers Are Co-Designed
Co-Designed for Stability
Each layer reduces the burden on the others:
- Mechanics reduce instability before it becomes a sensing problem
- Sensing reduces uncertainty before it becomes a control problem
- Reflex enforcement prevents unsafe commands from becoming hardware failures
This is what makes the system stable.