DEPLOYDatabase

Robot capabilities

What sensors do robots use?

Depends on the robot

There is no single sensor; the suite depends on the robot and its job. Robots use different sensors for different jobs, so there is no single answer. A Waymo robotaxi fuses lidar, cameras, and radar. A Tesla car uses cameras only, no lidar. A Figure humanoid uses onboard cameras and microphones. A DJI drone uses vision sensors and satellite positioning. A Roomba uses cliff, bump, and camera sensors. Each suite is attributed to its maker or press below.

Last verified Aug 19, 20247 sources

Find out instantly when it changes.

Verified changes only. We check the record and link every source. No marketing.

There is no one sensor a robot uses, because the right sensors depend entirely on what the robot has to do. The useful way to answer the question is to compare a handful of common machines, and that is exactly what the table below does, with every suite attributed to the maker who disclosed it or the press that reported it.

The sharpest contrast is in self-driving. A Waymo robotaxi uses sensor fusion: its 6th-generation Driver carries lidar, cameras, and radar working together, so it reads 3D shape, color and detail, and motion in bad weather all at once. A Tesla car takes the opposite bet, running its driver assistance on cameras only, with no lidar and no radar after Tesla removed them. Same broad goal, two very different sensor philosophies.

Other robots match their sensors to their world. A Figure humanoid perceives through onboard cameras and talks through microphones and speakers. A DJI drone pairs an omnidirectional vision system for obstacle sensing with satellite positioning to hold its place in the sky. A robot vacuum like a Roomba stays cheap and low: infrared cliff sensors keep it off the stairs, a bumper feels for walls, and newer models add a camera to map the room.

So the honest answer is that a robot's sensors are a design choice, tuned to its job and its budget. Everything here is dated and attributed, so you can check each suite against its source.

Sensor suites by robot type

There is no single answer, so here is the honest one: the sensors each common robot uses, with every suite attributed to its maker or the press that reported it, and dated to its source.

Sensor suites by robot type (claimed, attributed to the maker or press)
RobotSensor suiteWho (attributed)As ofSource

Waymo robotaxi (self-driving taxi)

Sensor fusion

Sensor fusion: lidar, cameras, and radar working together.

Waymo describes its 6th-generation Driver as carrying 13 cameras, 4 lidar, and 6 radar, plus external audio receivers, which work in unison so cameras read color and detail, lidar reads 3D shape and distance, and radar tracks motion in bad weather. This is the fusion approach, opposite to a camera-only design.

Waymo (6th-generation Waymo Driver)Aug 19, 2024

Figure 02 humanoid

Onboard vision plus audio

Onboard cameras plus microphones and speakers.

Figure says its Figure 02 humanoid perceives the world through 6 onboard RGB cameras feeding a vision language model, and it holds speech conversations through onboard microphones and speakers. The cameras are the primary perception, in the robot's head and body rather than a spinning roof unit.

Figure (Figure 02 announcement)Aug 6, 2024

DJI consumer drone

Vision plus GNSS

Vision sensors for obstacle sensing, infrared, plus satellite positioning.

DJI's Mini 4 Pro specifications list an omnidirectional binocular vision system for obstacle sensing, a downward 3D infrared sensor, and satellite positioning across GPS, Galileo, and BeiDou. The cameras let the drone see and avoid obstacles in every direction while GNSS holds its position in the air.

DJI (Mini 4 Pro specifications)Sep 25, 2023

Tesla car (Autopilot and FSD Supervised)

Cameras only

Cameras only: 8 cameras, no lidar, no radar.

Tesla runs its driver assistance on cameras alone. It removed radar from Model 3 and Model Y in 2021 and later dropped ultrasonic sensors, moving to a camera-based system it calls Tesla Vision, built on 8 cameras with no lidar. This is the vision-first bet, the opposite of Waymo's fusion suite.

Tesla (Tesla Vision), reported by The Drive and ElectrekMay 25, 2021 (as reported)

iRobot Roomba (robot vacuum)

Contact, cliff, and vision

Cliff and bump sensors, acoustic dirt detection, and a camera for mapping on newer models.

A Roomba uses infrared cliff sensors so it does not fall down stairs, a bumper that senses contact with obstacles, and acoustic dirt sensors that focus on dirtier spots. Newer models add an upward-facing camera and a floor-tracking sensor to build a map of the home through vSLAM, so it cleans in efficient rows rather than at random.

iRobot (Roomba), reported by IEEE Spectrum and WikipediaSep 6, 2018 (as reported)

Two philosophies of self-driving

The sharpest split is in autonomous driving. Each column is attributed to who holds the position: the fusion camp, the vision camp, and the honest bottom line that the sensors follow the job.

The fusion camp

Lidar, cameras, and radar together

Waymo fuses lidar, cameras, and radar so each sensor covers the others' blind spots. It reads 3D shape, color and detail, and motion in bad weather at once, and pays for it in cost and complexity.

The vision camp

Cameras only

Tesla runs on cameras alone, with no lidar and no radar, betting that neural networks can read the road from images the way a person does. It is cheaper and simpler, and it is the more contested bet.

Where it stands

The sensors follow the job

There is no single right suite. A humanoid carries cameras on its body, a drone pairs vision with satellite positioning, and a robot vacuum stays cheap with cliff and bump sensors. Each maker's choice is attributed and dated below.

How we know this

Every sensor suite on this page is attributed to the maker who disclosed it or the reputable press that reported it, with the primary source for each row. DEPLOY holds no structured sensor-suite fact class, so nothing here is a DEPLOY-verified spec. Here are the sources behind them.

Each suite is a maker disclosure or a press report, attributed to its source and dated, not a figure DEPLOY independently verifies. DEPLOY does not sell robots or software; it checks the record and links every source.

Common questions

What sensors does a self-driving car use?
It depends on the company. Waymo uses sensor fusion, combining lidar, cameras, and radar so each sensor covers the others’ weak spots. Tesla takes the opposite approach, using cameras only, with no lidar and no radar, and relying on neural networks to interpret the images. Both aim at autonomy, but Waymo bets on layered sensing and Tesla bets on vision alone. Each suite is attributed to its maker below.
Do Tesla cars use lidar?
No. Tesla runs its Autopilot and Full Self-Driving (Supervised) on cameras only. It removed radar from Model 3 and Model Y in 2021, later dropped ultrasonic sensors, and moved to a camera-based system it calls Tesla Vision, built on 8 cameras with no lidar. That is a deliberate choice, and it is the main way Tesla's sensing differs from Waymo's fusion suite.
What sensors do humanoid robots use?
Mostly onboard cameras, plus microphones, and internal sensors for balance and force. Figure AI says its Figure 02 humanoid perceives through 6 onboard RGB cameras feeding a vision model and talks through onboard microphones and speakers. Humanoids carry their sensors on the body rather than on a spinning roof unit, because they have to move through human spaces and manipulate objects up close. Suites vary by maker, and each is attributed below.
How do drones sense obstacles?
Consumer drones use vision sensors backed by satellite positioning. DJI's Mini 4 Pro specifications list an omnidirectional binocular vision system that lets the drone see and avoid obstacles in every direction, a downward infrared sensor, and satellite positioning across GPS, Galileo, and BeiDou to hold its position. The cameras handle avoidance while the satellite fix keeps the drone steady in the air.
What sensors are in a robot vacuum?
Simple, cheap ones tuned to a home. A Roomba uses infrared cliff sensors so it does not fall down stairs, a bumper that feels contact with walls and furniture, and acoustic dirt sensors that focus on dirtier spots. Newer models add an upward-facing camera and a floor-tracking sensor to map the home through vSLAM, so the robot cleans in orderly rows instead of bouncing around at random.

Keep reading

Ask the registry

Track this answer

Find out instantly when it changes.

Verified changes only. We check the record and link every source. No marketing.

Canonical URL: /what-sensors-do-robots-use · machine-readable mirror: /what-sensors-do-robots-use.md