How Do Robots Work? Sensors, Brains and Motors Explained, From Factory Arms to Humanoids

Robots have moved from science fiction to everyday reality. Factories around the world installed 542,076 industrial robots in 2024 alone, according to the International Federation of Robotics (IFR), and a new generation of humanoid robots is now being tested in warehouses and car plants. But how does a robot actually work? This guide explains the parts every robot shares, how industrial arms, collaborative robots and humanoids differ, where artificial intelligence fits in and what the numbers say about the robot boom.

What Is a Robot, Exactly?

A robot is a machine that can sense its surroundings, decide what to do and then act on the physical world, often without a human controlling each movement. That sense-think-act loop is what separates a robot from a simple machine such as a food mixer, which only does one thing when you press a button.

Robots come in many shapes. The most common by far is the industrial robot arm bolted to a factory floor. There are also wheeled robots that move goods around warehouses, robot vacuum cleaners in homes, surgical robots in operating theatres, drones, self-driving cars and, increasingly, humanoid robots built to resemble the human body.

What Are the Main Parts of a Robot?

Almost every robot, from a $300 vacuum to a factory arm costing tens of thousands of dollars, is built from the same five building blocks:

• Sensors: the robot’s eyes and ears. Cameras, lidar, ultrasonic sensors, force and torque sensors, microphones and encoders that measure how far each joint has turned
• Controller: the computer or “brain” that runs software, processes sensor data and sends commands
• Actuators: the muscles. Usually electric motors, sometimes hydraulic or pneumatic cylinders, that move joints or wheels
• End effector: the tool at the end of an arm, such as a gripper, suction cup, welding torch or paint sprayer
• Power supply: mains electricity for fixed robots or batteries for mobile ones

Engineers describe a robot arm by its degrees of freedom, the number of independent ways it can move. A typical six-axis industrial arm has six rotating joints, roughly matching the freedom of a human arm from shoulder to wrist, which lets it reach a point from almost any angle.

How Does a Robot Know Where to Move?

Traditional industrial robots are programmed, not intelligent. An engineer either writes code or uses a handheld “teach pendant” to guide the arm through a sequence of positions, which the controller records and repeats thousands of times a day with accuracy measured in fractions of a millimetre.

Behind the scenes, the controller solves maths problems called kinematics. Forward kinematics works out where the gripper will be if each joint is set to certain angles. Inverse kinematics does the reverse, calculating the joint angles needed to put the gripper at a target point. A feedback loop then compares the planned position with readings from the joint encoders and corrects any error many times a second.

Mobile robots add another layer called localisation and mapping. A warehouse robot or robot vacuum builds a map of its surroundings using cameras or lidar and works out where it is on that map, a technique known as SLAM (simultaneous localisation and mapping). The same idea powers self-driving cars, which fuse data from many sensors to plan a safe path. If you have ever wondered how a robot vacuum avoids your sofa, this is how.

How Does AI Change What Robots Can Do?

Classic robots excel at repetitive tasks in tightly controlled spaces but struggle when anything changes, such as a part arriving at a slightly different angle. Artificial intelligence is loosening that limit in three main ways:

• Computer vision: neural networks let robots recognise objects, read labels and spot defects, so they can pick items from a cluttered bin rather than a neat tray
• Learning from demonstration and simulation: robots can be trained in virtual worlds millions of times before trying a task for real, or learn by copying human demonstrations
• Language and reasoning models: newer “vision-language-action” models let a robot follow plain-English instructions such as “put the red cup in the sink”

These advances are why companies are pouring money into general-purpose robots. Even so, today’s AI-driven robots are slower and less reliable than people at many everyday tasks, and most real-world deployments still involve narrow, well-defined jobs. For background on the AI side, see our explainer on how AI is reshaping everyday life.

What Is the Difference Between Industrial Robots, Cobots and Humanoids?

Industrial robots are powerful, fast arms used for welding, painting, assembly and packing. Because they can injure people, they usually work inside safety cages or fenced cells.

Collaborative robots (cobots) are smaller, lighter arms designed to work alongside people. They use force sensors and speed limits so they stop if they bump into someone, and international standards such as ISO/TS 15066 set out how much contact force is acceptable. Cobots are popular with small manufacturers because they are cheaper and easier to reprogram.

Humanoid robots have a head, torso, two arms and usually two legs. The argument for the human shape is practical: our factories, warehouses and homes are built for human bodies, so a humanoid could in theory use the same stairs, tools and shelves. Well-known examples include Boston Dynamics’ electric Atlas, Agility Robotics’ Digit, Figure AI’s robots, Tesla’s Optimus prototypes and Chinese models such as Unitree’s G1, which was advertised from about $16,000 when it launched in 2024. Most humanoids are still in pilot projects rather than mass deployment.

How Many Robots Are There in the World?

The IFR’s World Robotics 2025 report gives the clearest picture of the industrial robot market:

• 542,076 industrial robots were installed worldwide in 2024, the second-highest annual total on record and more than double the figure of ten years earlier
• The global operational stock reached 4,663,698 robots, up 9% on the previous year
• China installed 295,045 robots in 2024, 54% of the world total, and has more than 2 million industrial robots in operation
• Japan, the United States, South Korea and Germany make up the rest of the top five markets

Car-making and electronics remain the biggest users, but logistics, food and drink, and metal products are growing quickly as robots become cheaper and easier to deploy.

Will Robots Take Our Jobs?

Robots tend to replace specific tasks rather than whole jobs, especially tasks that are dangerous, dirty or dull. Economists have found that heavy robot adoption can reduce employment and wages in some local manufacturing areas, while also raising productivity and creating jobs in programming, maintenance and integration.

Humanoids and AI-powered robots could widen the range of tasks that can be automated, which is why skills in maintaining, programming and supervising machines are in growing demand. Our guide to the skills you need in an AI economy looks at how workers can prepare.

How Can You Learn More or Get Started in Robotics?

You do not need a factory to explore robotics. Here are practical first steps:

• Start with a kit: beginner kits based on Arduino or Raspberry Pi let you build a small wheeled robot and learn sensors, motors and code
• Learn Python and basic maths: most robot software, including the open-source Robot Operating System (ROS 2), uses Python or C++, plus geometry and linear algebra
• Try simulation: free tools such as Gazebo let you test robot code on a virtual robot before touching hardware
• Join a competition: school and college programmes such as FIRST Robotics give hands-on team experience
• Look at cobot training: major robot makers offer short online courses aimed at technicians and engineers

What Are the Limits and Risks of Robots?

Robots remain expensive to install and integrate, and a robot that works perfectly in a demonstration video may struggle with the variety of real life. Batteries limit how long mobile and humanoid robots can run, and grasping soft or unfamiliar objects is still hard.

Safety and security matter too. A malfunctioning or hacked robot can cause physical harm, so manufacturers follow strict safety standards and businesses need to secure robot software like any other connected device. Privacy is another concern when home robots carry cameras and microphones, so check what data a device collects and where it is stored before you buy.

The Bottom Line

Every robot combines sensors, a controller, actuators and a tool to sense, decide and act. Industrial arms already do much of the world’s welding and assembly, cobots are bringing automation to smaller firms, and AI is pushing robots, including humanoids, into less predictable places. The technology is advancing fast, but for now the most useful robots are still the ones built to do one job very well.

Featured image: Japan-Mobility-Show-2025-RuinDig 0567, by RuinDig/Yuki Uchida / Wikimedia Commons, CC BY 4.0. Cropped, colour-graded and combined with MediaPavilion branding.

Sources

• International Federation of Robotics — Global Robot Demand in Factories Doubles Over 10 Years (September 2025).
• International Federation of Robotics — World Robotics 2025 Executive Summary: Industrial Robots (2025).
• ISO — ISO/TS 15066:2016 Robots and robotic devices: Collaborative robots (2016).
• Boston Dynamics — Atlas (accessed 2 November 2026).
• Unitree Robotics — Unitree G1 humanoid robot (accessed 2 November 2026).
• Open Robotics — ROS 2 documentation (accessed 2 November 2026).