How does millimeter wave radar count people around a vending machine?

A millimeter wave radar transmits a low-power radio signal and measures the reflections. From the reflections it calculates each person's distance, direction and speed, and tracks them as moving targets inside a defined detection zone. Software counts a target as an entry when it crosses into the zone and as an exit when it leaves, and shows how many targets are in the zone right now. In KioskForce's demo, the radar was mounted at 1.9 m, about the height of a KioskForce beverage vending machine, and the counts updated in real time as an engineer walked in and out.

Is radar people counting anonymous?

Yes. A radar does not capture an image. It measures reflected radio energy and outputs positions and movement, not faces, clothing or skin colour, so it cannot recognise or identify anyone. That is the difference from a camera, even one running anonymised analytics: there is no picture to leak, retain or misuse. Radar counting therefore works in places where cameras are not acceptable, such as restrooms, change rooms, locker rooms, clinics and schools.

Can a camera do the same job?

Where cameras are allowed, a camera with on-device analytics can count people and can also estimate things radar cannot, such as whether a person looked at the screen. But a camera captures images of people, which brings privacy obligations and is simply not possible in restrooms, change rooms and many healthcare settings. Radar counts traffic without capturing any image, works in the dark, and is not affected by lighting. For a pure traffic count, radar is the lower-risk sensor.

Is mmWave radar traffic counting available on KioskForce machines?

It is a technical preview. KioskForce has built and demonstrated a working radar traffic counter. Because every KioskForce machine is designed to specification, radar counting can be added to a custom vending machine, kiosk or smart locker project on request. The mounting position, detection zone and reporting are then designed around the site. Contact KioskForce with the site type and what you want to measure.

Millimeter Wave Radar Traffic Counting

Q: What does the radar measure?

The KioskForce traffic monitoring demo reports four values: entries (people who walked into the detection zone), exits (people who walked out), current (people standing in the zone now), and monitoring duration. It also lists the targets currently detected and keeps a system log. Combined with the machine's own sales data, these counts give conversion rate (purchases divided by people who passed), peak hours, and how often people stop in front of the machine without buying.

KioskForce has demonstrated millimeter wave (mmWave) radar as an anonymous foot-traffic counter for vending machines. A radar module mounted at 1.9 m, about the height of a KioskForce beverage vending machine, counts the people who enter and leave the area in front of the machine and shows how many are standing there now. Radar measures distance and movement, not appearance. It records no images and cannot recognise a face, so it can count traffic in restrooms, change rooms, clinics and other places where a camera cannot be installed. Divide sales by traffic and you get a machine’s conversion rate. That tells you whether a weak location has too few people walking past or loses the people who do. This is a technical preview: KioskForce builds it into custom projects on request.

Anonymous foot traffic, measured at the machine

A vending machine’s sales report tells you what sold. It does not tell you how many people walked past and didn’t buy.

That missing number is the difference between two very different problems:

Low traffic. Few people pass the machine. Moving it will help more than changing the products.
Low conversion. Plenty of people pass the machine, but few buy. Fix the price, the product mix, the screen, or the payment options.

KioskForce has been testing millimeter wave (mmWave) radar as the sensor that measures this number. It counts people without a camera.

Technical preview: a KioskForce engineer walks in and out of the radar's detection zone while the traffic counts update in real time.

What the demo shows

The demo in the video above is our own engineering prototype:

The radar is mounted at 1.9 m. That is about the height of a KioskForce beverage vending machine, so the sensor sees the area in front of the machine from the same position it would have on the cabinet.
It connects over USB. For the demo, the radar plugs into a laptop running the Millimeter Wave Radar Traffic Monitoring System. The operator presses Refresh, selects the radar’s port, presses Connect, then Start Monitoring.
The software reports four numbers:
- Entries: people who walked into the detection zone
- Exits: people who walked out of it
- Current: people standing in the zone right now
- Duration: how long monitoring has run
The counts update in real time. When our engineer walks into the zone, the entry count goes up and Current shows one person. When the engineer walks out, Current returns to zero.

The screen also lists the targets currently detected and keeps a system log of connection and monitoring events.

Laptop running the KioskForce Millimeter Wave Radar Traffic Monitoring System, showing entries, exits, current count and monitoring duration — The traffic monitoring software: entries, exits, people currently in the zone, and monitoring duration.

Why radar instead of a camera

A camera captures images of people. Even with anonymised analytics running on the device, there is a picture to secure and a privacy notice to justify, and in some places you cannot install a camera at all. We covered where that line sits in Australia’s facial recognition guidance for vending cameras.

A radar never captures an image. It measures reflected radio energy and turns it into positions and movement.

	mmWave radar	Camera with analytics
What it captures	Distance, direction and speed of moving targets	Images of people
Can it identify a person?	No: there is no image of a face	Yes, if the software or footage allows it
Restrooms, change rooms, clinics	Usable	Generally not acceptable
Darkness and changing light	Unaffected	Degrades without lighting
Behind the machine’s panel	Can sit behind a plastic cover	Needs a clear line of sight through a lens opening
Attention metrics (did they look at the screen?)	No	Yes
Traffic count	Yes	Yes

If all you need is a traffic count, radar is the lower-risk sensor. Choose a camera only if you also need attention or audience analytics, and only in places where a camera is appropriate.

Where radar counting fits

The use case that started this work was restrooms: anonymous traffic tracking where a camera-based system cannot be used. The same logic applies wherever a machine sits in a private or sensitive space:

Washrooms and restrooms. Wall-mount vending machines for sanitary products, sexual health and hygiene items.
Change rooms and locker rooms. Gyms, pools, sports clubs and staff amenities.
Healthcare and clinics. Waiting areas and corridors where patient privacy rules out cameras.
Schools and campuses. Machines placed near students.
Site selection before a machine goes in. Mount a radar at a candidate location, measure traffic for a few weeks, and compare sites before committing a machine.
Screen advertising. The number of people in front of a machine is the audience for its advertising screen. A radar count can back an impression report without a camera.

What to plan for

Radar is a precise sensor, but it is not a drop-in one. These are the decisions we make with you for each project:

Mounting height and angle. The demo used 1.9 m. The right position depends on the cabinet and on the corridor or room in front of it.
Detection zone. Draw the zone around the area you want to count, for example the area in front of the machine and not the whole corridor. A tighter zone gives a more meaningful count.
Surroundings. Metal surfaces and crowded spaces affect radar reflections. Every site needs testing on the day of installation.
Local radio rules. mmWave modules operate in licence-exempt radio bands, but the permitted band varies by country. The module has to suit the market the machine ships to.
Reporting. In the demo, the radar connects to a laptop over USB. In a custom build, it connects to the machine’s own controller, so traffic can be reported next to sales from the same machine.

Status: technical preview

Radar traffic counting is a technical preview. We have built and demonstrated it, and we can build it into a custom vending machine, kiosk or smart locker project. It is not a standard option on a catalogue machine, because KioskForce does not sell catalogue machines. Each system is designed to the customer’s specification.

If you run machines in places where a camera is not an option and you want to know how many people pass them, tell us about the site.