Four people are sitting in a meeting room, everyone is watching the slides, nobody gets up. Ten minutes in, the lights switch off, someone waves both arms, and they come back on. That moment sums up the occupancy sensor vs motion sensor question: a motion sensor reports that something moved, while an occupancy sensor is supposed to report that someone is there, even when that person is only reading, typing or listening. The two devices often look alike and hang in the same spot on the ceiling, but they answer two different questions.
For choosing hardware, that means the first decision is not about range or price. It is about which behavior in the room has to be detected. A hallway that people walk through has different needs than a workstation where someone works almost motionless for two hours. Settle that first, and you skip the endless tweaking of sensitivity and delays that only hides symptoms.
How does a motion sensor work?
Most motion sensors use a PIR sensor. PIR stands for passive infrared: the sensor emits nothing and simply receives the heat radiation every body gives off. At its core is a pyroelectric element that only responds to changes in incoming radiation. A lens in front of it, usually a Fresnel lens with many segments, splits the room into separate detection zones. When a person walks from one zone into the next, the heat pattern on the element shifts and the sensor fires.
That leads to an important property. A PIR sensor does not measure temperature and does not recognize people as such. It recognizes change. It responds most strongly when someone walks across the zones, past the sensor rather than toward it. A person heading straight at the sensor crosses fewer zones and produces a weaker signal.
After it triggers, a motion sensor keeps its output active for a set period, the time delay. Every new movement restarts the clock. Once it runs out with nothing happening, the sensor switches off or reports no motion. That is exactly right for hallway lighting, stairwells or an after-hours alarm, because every situation that matters there starts with movement.
How does an occupancy sensor detect people?
An occupancy sensor, often sold as a presence sensor, has a different goal: it should reflect the state “room in use” with as few gaps as possible. In conventional building automation, that is frequently a PIR device as well, but with a much finer zone pattern and higher sensitivity. That lets it pick up small movements such as typing, turning a page or a turn of the head. Many also measure light level.
In the US, you will also see the term vacancy sensor. It uses the same detection hardware but behaves differently: you switch the lights on by hand, and the sensor only switches them off once the room has been empty for the set time. That saves energy in rooms where daylight is often enough.
Newer presence sensors go further and use other sensing principles: millimeter-wave radar, thermopile sensors that measure heat directly, or image sensors with on-device analysis. What they share is that they deliver a state rather than a single event: occupied or vacant, and on some devices the number of people in a defined area. Behind that state there is always a decision rule. The device combines readings, holds the state for a while and only switches once there is enough evidence for vacant.
The label alone guarantees nothing. A device sold as an occupancy sensor may be a sensitive PIR unit with a long delay, so check the documentation for the exact variant and ask which physical signal it actually responds to.
Presence sensor vs motion sensor: why do motion sensors miss people sitting still?
The reason lies in the sensing principle itself. A PIR element responds to changes in heat radiation. Someone sitting quietly at a desk still gives off heat, but the heat pattern on the sensor barely changes. No change means no signal, and once the time delay runs out, the motion sensor treats the room as empty even though someone is sitting in it.
How quickly that happens depends on distance, furniture and activity. Small hand movements barely register near the edge of the coverage area, monitors hide the hands that are still moving, and while someone reads a document or sits through a video call on mute, they hardly move at all.
The usual reflex is to turn up the sensitivity. That only helps up to a point, because a PIR sensor cannot report information that is not physically there, and false triggers from radiator air, sunlight or passersby increase. A longer time delay bridges quiet periods but adds no measurement either; the room just stays occupied longer after everyone has left.
If a room is regularly used by people working quietly, this is not a settings problem. It is a sign that a plain motion sensor uses the wrong sensing principle for the job.
What is presence detection, and which technologies use it?
Presence detection means recognizing that a person is in a space, whether or not that person is moving right now. High-resolution PIR is the classic approach: many small zones so minor movements still register. It is cheap and efficient but still needs some movement.
An mmWave presence sensor transmits radio waves in the millimeter-wave range and analyzes the reflections. That lets it detect movements too small for a PIR sensor. Radar has its own quirks, though: the waves pass through thin walls, doors and glass, so movement in the next room or the hallway can show up as presence. Manufacturer documentation also lists fans, curtains and plants as typical sources of interference and points out blind spots. Some devices only switch on their radar after a PIR sensor has detected movement, then hold the occupancy state from there.
A thermopile measures heat radiation directly and so produces a reading that persists while a person sits still. That makes it attractive for individual workstations, which is why desk sensors often combine PIR and thermopile. Its limits are resolution and other heat sources in its field of view.
The Milesight VS121 takes a third route. It uses an image sensor with on-device AI analysis and detects people in up to 16 freely defined regions. According to the manufacturer, no image data leaves the device, only counts and states are transmitted, and the stated accuracy is up to 95 percent. Here too, a seat the sensor cannot see will not be detected.
When is a motion sensor good enough?
A motion sensor is the right choice wherever the relevant situation reliably starts with movement and people do not linger. Hallways, stairwells, storage aisles, basements and utility rooms all fit. Anyone there is walking, and that is exactly what a PIR sensor is built for. It also suits after-hours security, where every event is meaningful.
Motion events also work for rough statistics, such as whether a room was used at all on a given day. They do not tell you how many hours a workstation was actually occupied.
PIR sensors also need very little power, so wireless motion sensors often run for years on a battery, and they are cheaper and less prone to interference from neighboring rooms. A motion sensor is not an inferior solution. It is the right solution to a different question.
The rule of thumb: if the longest quiet stretch in the room is shorter than a time delay you can accept without annoying anyone, a motion sensor is enough. If people sit, read or listen there for long periods, you need a real presence sensor.
What time delay should an occupancy sensor use?
The time delay is how long a sensor keeps reporting occupied after the last detection. It covers gaps in detection, but it also stretches the time until a room that has really been vacated counts as empty. There is no universal ideal value, because both sides of that trade-off depend on what the sensor controls.
If the sensor controls lighting, a delay that is too short is mainly annoying: the lights go off while someone is working. If it controls heating, ventilation or air conditioning, energy weighs more heavily, and an unnecessarily long delay keeps equipment running in empty rooms. For utilization analysis, a long delay inflates every use.
A sensible approach is to start short and only extend the delay until typical quiet activities no longer cause false shutoffs. If you need very long delays to bridge the same detection gaps over and over, the problem is not the timer but the mounting position or the sensing principle. A presence sensor that genuinely detects people sitting still gets by with much shorter delays than a PIR unit.
A common mistake is stacking hold times. If the device has its own delay and the building controller or platform adds a second one, the two add up. Decide where the timer runs and document which event restarts it.
Where should you mount an occupancy sensor in an office?
Mounting position often matters more than the technology. The basic rule for any occupancy sensor in an office: it has to see the places where people sit, not just the door they walk through. A unit above the entrance misses the person working quietly at the far end of the table.
For ceiling-mounted PIR units, the best spot is usually centered above the seating area, so small movements at the table fall into several zones. Make sure the coverage area does not reach into the hallway through open doors or glass walls, or colleagues walking past will trigger it. Radiators, air vents and windows with direct sun do not belong in the coverage area.
Radar units need extra care around walls and doors because their waves pass through thin materials. Some manufacturer guides recommend closing meeting room doors after use and advise against very small rooms such as restrooms.
The Milesight VS121 is designed for ceiling mounting at 2.5 to 4 meters. With a field of view of 190 degrees horizontal and 112 degrees vertical, the manufacturer states it covers up to 127.84 square meters at a 4-meter mounting height. Its configurable regions let you define which zones count as desks or a meeting table, so walkways are not counted.
Wireless installation brings its own checks. The VS121 communicates over LoRaWAN in the EU868 band and runs in Class C, so it is always ready to receive. It is powered over USB-C at 5 volts, which means it is not a battery device and you need a power source on the ceiling. According to the manufacturer, it can buffer and resend up to 3,000 packets if the link to the gateway drops briefly, and it is configured through its own Wi-Fi access point. LoRaWAN is built for small packets, not instant switching, so it suits utilization data and demand-driven HVAC better than lighting. Track the age of the last message, too: missing data does not prove a room is empty.
Whatever the technology, test at the real seats, especially edge seats and spots behind monitors or shelves: read, type, join a video call, just listen. Then leave the room and time how long it takes to report vacant. Check the empty room too: if the sensor reports presence with nobody there, the boundaries or alignment are not right yet.
PIR vs radar vs thermopile: which presence sensor is most reliable?
No technology is the most reliable in every room. Each measures something different and fails in its own way. The table below compares the common options by whether they can, in principle, detect a person sitting still, and by what typically throws them off.
| Technology | What it measures | Person sitting still | Typical interference |
|---|---|---|---|
| PIR motion sensor | Change in heat radiation between zones | usually not, only with visible movement | Radiator air, sunlight, furniture blocking the view |
| PIR occupancy sensor | Finer zones, higher sensitivity | with small movements such as typing | Same as PIR, plus false triggers from higher sensitivity |
| mmWave radar presence sensor | Reflected radio waves, even tiny movements | in principle yes, depending on device and distance | Fans, curtains, plants, movement behind thin walls or glass |
| Thermopile | Heat radiation directly, as a temperature value | yes, as long as the body is in view | Other heat sources in view, low resolution |
| Image sensor with AI analysis | People in the image of a defined area | yes, if the person is in the field of view | Hidden seats, wrong mounting height |
Read the table as a map of failure modes rather than a ranking. PIR is reliable for movement and unreliable for stillness. Radar catches the smallest movements but needs a room whose boundaries it respects. Thermopile handles stillness at a single desk well but struggles with larger areas and other heat sources. An image sensor sees everyone in its field of view but nothing outside it. The most reliable presence sensor is the one whose weakness does not occur in your room.
Where can you buy occupancy and motion sensors?
Start with the behavior that has to be detected. If people only pass through, a PIR motion sensor is enough, and it is frugal, affordable and robust. If they sit quietly for long periods, you need a presence sensor whose sensing principle picks up small movements or none at all, meaning high-resolution PIR, radar, thermopile or an image sensor. Then check the interference sources of that technology in the actual room, from radiator air to thin walls to hidden seats. Plan the mounting position above the seats rather than over the door, keep the time delay as short as possible and let it run in only one place. For wireless devices, clarify power supply, reporting behavior and gateway coverage in advance. Work through these points and the occupancy sensor vs motion sensor question turns into a clear requirement for your room. If you are looking for a suitable device, you will find our models in the movement and presence sensors category.