Wi‑Fi sensing turns ordinary Wi‑Fi signals into tools for detecting people, movement, and even identity—without cameras or wearables. It’s not about tracking devices or securing networks. Instead, it’s about how the radio waves themselves reveal what’s happening in a space. Let’s see how this works in practice.

Wi‑Fi signals travel through the air, bouncing off walls, furniture, and people. When someone moves—even slightly—they disturb these signals. The way signals scatter and reflect changes with every motion or presence. Wi‑Fi sensing measures these disturbances to detect what’s happening in the environment.

Channel State Information (CSI) captures detailed changes in how Wi‑Fi signals travel. Received Signal Strength Indicator (RSSI) measures overall signal power. Signal phase tracks the timing of the wave. These measurements form the raw data for Wi‑Fi sensing.

Wi‑Fi devices record CSI, RSSI, and phase data over time. Signal processing algorithms then extract patterns or detect anomalies. Machine learning models interpret these patterns—classifying them as presence, motion, or even specific gestures. The pipeline turns noisy radio data into actionable insights.

Even a stationary person subtly disturbs Wi‑Fi signals compared to an empty room. Algorithms look for persistent, low-level changes in CSI or RSSI to infer presence. However, false positives may occur—pets or moving objects can mimic human presence if not properly accounted for.

Walking, waving, or moving limbs create distinct, time-varying changes in CSI and phase. Algorithms can classify these patterns to recognize gestures—like waving or pointing. The accuracy depends on the environment, device placement, and the quality of training data used by the models.

Wi‑Fi sensing can estimate body pose—such as standing, sitting, or lying down—by analyzing complex CSI patterns. With high-resolution data, it can even track breathing or heartbeats by detecting minute signal fluctuations. These advanced uses require robust algorithms and clean signal capture.

Each person’s gait—their unique way of walking—modifies Wi‑Fi signals in characteristic ways. Machine learning can recognize these patterns to identify individuals. However, accuracy can drop if someone changes shoes, carries objects, or wears bulky clothing.

Walls, furniture, and other obstacles can block or unpredictably reflect Wi‑Fi signals. Multiple people or pets can confuse algorithms. Device placement and choice of Wi‑Fi frequency band (2.4 GHz vs 5 GHz) also affect sensitivity and range—there’s no one-size-fits-all setup.

Wi‑Fi sensing can detect presence or movement—even through walls—without people’s knowledge. Unlike cameras, it works in darkness and doesn’t need a direct line of sight. Responsible deployment demands transparency, user consent, and safeguards to prevent misuse.

Pets, moving fans, or robotic vacuums can be mistaken for humans. Rearranging furniture or changing the environment can degrade accuracy. For reliable results, training data must match real-world conditions where the system is deployed.

Wi‑Fi sensing uses CSI, RSSI, and phase to detect presence, motion, gestures, pose, breathing, and even identity—no cameras or wearables required. Smart algorithms extract meaning from signal disturbances, but accuracy and privacy depend on careful design and deployment.