The simple idea
A room that is empty does not need to be lit, and it rarely needs to be heated or cooled to full comfort. People, however, forget to switch things off, and fixed schedules cannot predict a meeting that runs late or a classroom that is unused for a week. Occupancy sensing replaces guesswork with a signal: is anyone here right now, and roughly how many?
With that signal, a lighting circuit can turn off a few minutes after the last person leaves, and a heating or ventilation system can relax to an unoccupied setting until someone returns.
How presence is detected
Different sensors answer slightly different questions, and each has blind spots.
- Passive infrared (PIR). Detects changes in infrared radiation caused by moving warm bodies. It is inexpensive and common in lighting controls. Its weakness is that it needs movement; a person sitting still at a desk may be missed.
- Ultrasonic and microwave sensors. Emit a signal and watch for changes in the echo. They can pick up smaller movements than PIR and see around some obstacles, but can sometimes be triggered by air movement or activity outside the room. Motion detection in connected devices covers these methods in more detail.
- Dual-technology sensors. Combine two methods and require agreement, which reduces both false triggers and false “empty” decisions.
- Carbon dioxide sensors. People exhale carbon dioxide, so rising levels suggest occupancy, and the level gives a loose indication of how many people are present. The response is slow, which suits ventilation more than lighting.
- Door and desk sensors. Simple contact or pressure sensors track entry or use of a particular space.
- Camera-based or people-counting sensors. These can count people accurately in some setups but raise stronger privacy questions, covered below.
- Wi-Fi or Bluetooth device counts. Some systems infer presence from the number of connected devices. This is indirect: people carry zero, one or several devices.
Lighting: the easier case
Lighting responds instantly, so presence-based control is relatively straightforward. Common approaches are:
- Automatic on, automatic off. Lights turn on when someone enters and off after a time delay once movement stops.
- Manual on, automatic off. A person switches the lights on, and the sensor turns them off. This avoids lights coming on when someone only passes a doorway, and is often considered a better fit for rooms with windows.
- Zoning and dimming. In large open areas, only the zone near people is fully lit, while others dim to a low level.
- Daylight combination. Presence control pairs well with daylight sensing so artificial light is not added when a room is bright.
The delay is a design choice. Too short, and lights switch off on a quiet worker. Too long, and the savings shrink. Rooms where people are often still, such as offices, usually need a longer delay than corridors.
Heating, cooling and ventilation: the slower case
Heating and cooling respond slowly because rooms and structures hold heat. That changes how presence data is used.
Rather than switching off, the system typically applies a setback: a wider acceptable temperature range when the space is empty. Because a cold room takes time to warm, the system may need a start time based on a schedule or an earlier trigger, such as a building entrance event, rather than waiting for someone to sit down and feel cold. Smart thermostats use a simple version of this with “home” and “away” modes.
Ventilation can respond to occupancy directly. In spaces with changing crowds, such as meeting rooms or lecture halls, the system can lower the airflow when few people are present and raise it as the room fills. Combining presence with air quality sensors helps keep air fresh while avoiding unnecessary fan power.
An example of the effect
Here is a hypothetical office meeting room with lighting drawing 200 W. If it is occupied 3 hours of an 11-hour working day but lit the whole time, it wastes 8 hours x 200 W = 1,600 Wh, or 1.6 kWh, per day. Over 250 working days, that is 400 kWh. With presence control the waste would shrink to the delay periods. The numbers are invented, but they show how cheaply unneeded hours accumulate. Real savings vary with how often the space is used, so a short measurement beforehand is wise.
What can go wrong
- Poor placement. A sensor behind a partition or aimed at a heating vent may report wrong data. Careful sensor placement and testing in the actual room matter more than the sensor model.
- False “empty” readings. Lights switching off on a seated person is the best-known annoyance. People tend to disable systems that do this, which erases savings.
- False “occupied” readings. Pets, moving curtains or hot air can keep a room active.
- Conflicting controls. A presence sensor, a timer and a manual switch need clear priority rules.
- Safety. Lighting in stairwells and escape routes needs special care, and requirements vary by location. Follow local rules and qualified advice.
Privacy considerations
A presence signal shows when people are in a space, and in some set-ups it can be linked to individuals. For energy control, a simple “occupied or not” value, or an anonymous count, is usually enough. Avoid collecting images or identifying information when a basic sensor will do, tell people what the sensors do, and keep records no longer than needed. These principles are discussed further under workplace monitoring.