> ## Documentation Index > Fetch the complete documentation index at: https://docs.waveyhq.dev/llms.txt > Use this file to discover all available pages before exploring further. # Security and Intrusion Detection via WiFi CSI > Detect motion through walls without cameras, wearables, or light. WiFi CSI enables device-free, darkness-invariant zone monitoring using existing radio signals. WiFi CSI detects intruders by sensing how a human body disturbs the radio channel. It works in total darkness, through drywall partitions, and without requiring the target to carry any device. This makes it a practical complement to cameras and PIR sensors for zone-level security. ## NLOS motion through building materials WiFi at 2.4 and 5 GHz penetrates drywall, plaster, and wooden doors. An intruder moving in an adjacent room perturbs multipath paths that traverse the wall. Amplitude shifts propagate to the receiver even without line of sight. This is zone-level detection: "motion occurred in this area," not "person at coordinates (x, y)." Concrete, brick, and metal-shielded walls attenuate heavily. Interior drywall partitions are the practical case. The [detection ladder](/concepts/detection-ladder) places through-wall localization at rung 10: partial motion detection is achievable, but precise reconstruction is not. ## Darkness and obstacle invariance CSI does not depend on ambient light. WiVi (CVPR Workshop 2019) demonstrated this quantitatively: a vision classifier dropped from 95% to 46.67% accuracy with lights off, while a WiFi CSI classifier held at 95.83%. Their system disables the vision module when brightness falls below threshold and falls back to WiFi-only. Decision-level fusion, not feature fusion, means one modality failing does not corrupt the other. A PIR sensor in a dark corner may miss an intruder crawling below its cone. A camera needs illumination or infrared supplementation. WiFi sensing reads the same signal at midnight and noon. Coverage is whole-room from a node pair, not a single cone. Obstacles that block line of sight to a camera may still allow radio paths to reach the receiver. ## Device-free vs MAC tracking MAC-address security tracks phones and laptops. This is useless when the intruder carries nothing, or when you want to detect any body regardless of device ownership. CSI sensing detects the physical perturbation of a human body on the channel itself. This also means CSI cannot be defeated by leaving a phone behind. The body is the signal. ## What you learn and what you do not | Signal | Available | Not available | | ---------------------- | --------- | ---------------------------------------------------------------------- | | Motion in a zone | Yes | — | | Approximate timing | Yes | — | | Identity | — | No (research: 91% gait ID among 20 enrolled subjects; not Wavey scope) | | Appearance | — | No | | Precise position | — | No (without dense node grid) | | Through concrete walls | — | No | Wavey baselines the protected space during known-empty periods and flags [motion](/use-cases/motion-activity-detection) and [presence](/use-cases/presence-detection) that violate the baseline. Pair with conventional alarms for verification: CSI tells you something moved, not who. Motion event semantics and activity classification Through-wall feasibility and capability tiers Deploy your first Wavey nodes