How does a time-of-flight sensor work?

It sends out light, usually invisible infrared, and measures how long the light takes to come back. Light travels about 30 cm in a nanosecond, so the round-trip time gives the distance.

Why do time-of-flight sensors struggle with glass?

Light passes through glass, so the sensor may see what is behind it, or only a faint echo from the glass itself. Mirrors send the light elsewhere. That is why robots often add sound-based sensors, which bounce off glass.

What is a time-of-flight sensor? Measuring distance with light

In short. A time-of-flight sensor sends out light and times its return: light covers about 30 cm per nanosecond, so a round trip of a few nanoseconds means a distance of tens of centimetres. Some sensors measure the delay directly; others measure how a modulated light wave has shifted. Small multi-zone sensors measure a grid of distances at once. They are compact, light and frugal with power, which is why robots love them, but glass, mirrors, very dark surfaces and bright sunlight can fool them.

Timing light

Light travels at about 300,000 km per second: roughly 30 cm in a billionth of a second. A sensor that sends a flash and catches its echo can turn the round-trip time into a distance, halving it because the light went there and back. Measuring times that short is the hard part, and there are two ways to do it.

Direct time of flight. The sensor sends short pulses and times when individual returns arrive, with very fast light detectors.
Indirect time of flight. The sensor sends continuously modulated light and measures how far the returning wave has shifted in phase. A survey of these “lock-in” cameras notes that they offer neither higher resolution nor a longer unambiguous range than other depth systems, but give depth and brightness together at high frame rates, in a compact, light and low-power package.

From one distance to a grid

The simplest sensors return one distance. Multi-zone sensors split their view into a small grid, for example 8 by 8 zones, and return a distance for each, many times a second. That is coarse next to a camera, but it is real depth, measured directly, with almost no computing needed: ideal for a small robot deciding whether there is room ahead.

What fools them

Glass. Light passes through, so the sensor may see the wall behind a glass door, or just a faint echo from the glass.
Mirrors. They send the light elsewhere, and the sensor measures a room that is not there.
Very dark surfaces. Less light comes back, so measurements get noisier.
Sunlight. It contains infrared too, which adds noise.
Sheer fabric. A thin curtain can return a weak echo in front of a stronger one from the wall behind it, a pattern that can also help to recognise it.

That is why robots combine sensors: why glass and mirrors confuse robots.

Where they are used

In phones, to focus the camera quickly and sense when you hold the phone to your ear. In robots and drones, to keep a distance from walls, measure the height above the floor and find a place to land. In buildings, to sense whether someone is present.

How Alkyon is designed to use them

Alkyon, our concept for a flying home companion, is designed with five depth sensors of 64 zones each, around it and below it, alongside a camera and sound-based sensors for glass and mirrors. The downward one measures a landing spot and its height in one frame. It is a concept in development, for Europe only: how indoor drones navigate.

Source, read 4 November 2026: Foix, Alenyà and Torras, lock-in time-of-flight (ToF) cameras: a survey, IEEE Sensors Journal, 2011. General information.