What Is Ambient Intelligence?

Ambient intelligence describes a class of systems that infer the state of a physical environment from distributed, mostly non-imaging signals — radio-frequency variation, motion, contact events, device presence — and turn that inference into a small number of meaningful states.

The defining characteristic is data minimization by construction. The system is designed to produce 'occupied', 'activity stopped' or 'unusual for this hour', not a recording that later has to be governed.

Components

A working ambient intelligence deployment usually has four layers.

• Sensing: the mix of modalities appropriate to the environment
• Fusion: combining signals so each compensates for the others' blind spots
• Baselining: learning what normal looks like for this specific space
• Escalation: routing only meaningful deviations to a human