What Is Robotics, Anyway? Core Concepts and Real-World Use Cases

Imagine you’re making coffee on a Saturday morning, and a small disk-shaped machine hums past your feet, carefully navigating around the couch legs and the cat sleeping on the rug before returning to its charging station to empty the dust it just collected. Later that day, you order a package, and a small wheeled robot pulls up to your apartment door to drop it off, beeping politely to let you know it’s there. That night, you watch a documentary about surgeons using tiny robotic arms to perform heart surgery through incisions no larger than a keyhole, with patients recovering in days instead of weeks. These machines could not look more different from one another, yet they all share the same core DNA: they are robots, built on the same three foundational subsystems that power every autonomous machine on Earth today.

If you’ve ever assumed robotics is only for teams of PhDs in lab coats, or that robots are limited to the humanoid androids of science fiction, this chapter will change your perspective. We will break down exactly what robotics is, demystify the three core capabilities that all robots share, explore real-world use cases across every industry, and outline the step-by-step learning path you will follow to build your own functional robot by the end of this book. By the end of this chapter, you will be able to look at any machine and identify its perception, planning, and control subsystems, and understand how they work together to make it autonomous.

---

What Exactly Is a Robot?

Before we dive into the technical details, let’s start with a clear, practical definition that applies to every robot, from a $20 toy car to a $2 million surgical system: > <concept>A robot is a physical system that integrates three core capabilities to interact with and manipulate its environment autonomously, without continuous human input.</concept>

Notice that this definition does not require a robot to have a humanoid shape, a voice, or artificial general intelligence. It does not even require a robot to be able to “think” in the way humans do. All it requires is the integration of three subsystems: 1. Perception: Sensing and interpreting the robot’s environment and its own internal state. 2. Planning: Deciding what actions to take to achieve a predefined goal. 3. Control: Executing those actions precisely and reliably, even in the face of unexpected disturbances.

To make this definition concrete, let’s use a simple example most of us have interacted with: an automatic sliding door. Is it a robot? By our definition, yes:

• Its perception subsystem is the infrared motion sensor that detects when someone is approaching.
• Its planning subsystem is the simple logic circuit that

That simple automatic door, then, is the perfect microcosm of all robotics: its infrared sensor is its perception subsystem, its control circuit is its planning subsystem, and its motorized hinges are its control subsystem, all working in a tight loop to achieve the single goal of “let people pass through when they approach.” If you can break down that door into its three core components, you already understand the foundational architecture of every robot on the planet, from the smallest toy to the largest Mars rover. Let’s expand this definition and break down each subsystem in detail, with concrete examples that will help you spot the “robot DNA” in any machine you encounter.

---

The Three Pillars of All Robotics: Perception, Planning, Control

Every autonomous robot, no matter how simple or complex, is built on three interdependent subsystems that form a continuous feedback loop. We call these the *perception-planning-control loop*, and it is the universal architecture that has powered every successful robot since the first autonomous machine was built in the 1960s. To understand how this loop works, let’s use a more complex example you likely interact with regularly: a modern robot vacuum cleaner.

1. Perception: The Robot’s Senses

Perception is the subsystem that collects data about the robot’s environment and its own internal state, then processes that data to extract meaningful information. Think of perception as the robot’s eyes, ears, skin, and inner ear all rolled into one: it tells the robot where it is, what’s around it, and how it’s moving.

Robots use a wide range of sensors to gather perception data, each suited to different tasks:

• **Vision sensors**: Cameras (monocular, stereo, or depth cameras) that capture visual data to identify objects, read text, or detect obstacles. A robot vacuum uses a downward-facing camera to detect dark, soft surfaces like carpets, while an autonomous vehicle uses forward-facing cameras to read stop signs and traffic lights.
• **LiDAR (Light Detection and Ranging)**: Sensors that emit laser pulses and measure the time it takes for the light to bounce back, creating a precise 3D map of the surrounding environment. Most mid-range robot vacuums use LiDAR to map your home’s layout, while self-driving cars use long-range LiDAR to detect vehicles and pedestrians hundreds of meters away.
• **Inertial Measurement Units (IMUs)**: Sensors that measure acceleration, angular velocity, and orientation, similar to the sensor in your phone that detects when you rotate the screen. Robots use IMUs to detect if they’ve been bumped, if they’re tilted (like when a robot vacuum drives over a thick rug), or how fast they’re turning.
• **Proprioceptive sensors**: Sensors that measure the robot’s own internal state, rather than the external environment. The most common example is an *encoder*, a small sensor attached to a robot’s motor that counts how many times the wheel or joint has rotated. For a robot vacuum, encoders tell it exactly how far it has traveled, even if its wheels slip on a slippery floor. For a robotic arm, joint encoders tell the robot the exact angle of each joint, so it knows the position of its end effector (the “hand” at the end of the arm).
• **Environmental sensors**: Specialized sensors that measure properties of the environment, like temperature, humidity, gas concentration, or pressure. A firefighting robot uses gas sensors to detect toxic fumes, while an agricultural robot uses soil moisture sensors to decide when to water crops.

To make this concrete, let’s look at the perception pipeline of a typical robot vacuum: 1. The LiDAR spins 360 degrees per second, emitting thousands of laser pulses per second to measure the distance to every wall, piece of furniture, and obstacle in the room. 2. The downward-facing camera takes photos of the floor to detect dark carpets, stains, or small objects like socks that should not be vacuumed. 3. The IMU measures if the robot is tilting, to detect if it’s driving over a step or threshold. 4. The wheel encoders count how many rotations each wheel has made, to calculate how far the robot has traveled since it started cleaning. 5. All of this data is combined and processed by an on-board computer to create a real-time map of the robot’s location and the surrounding environment.

2. Planning: The Robot’s Brain

If perception is the robot’s senses, planning is its brain: the subsystem that takes the raw data from perception, combines it with pre-defined goals, and decides what actions to take next. Planning happens at multiple levels, from high-level task scheduling to low-level motion decisions:

• **Task-level planning**: The highest level of planning that decides *what* the robot should do overall. For a robot vacuum, this might be the schedule: “Clean the living room on Monday, Wednesday, and Friday, then return to the charging station when the battery is below 20%.”
• **Path planning**: The mid-level planning that decides *how* to get from point A to point B. For a robot vacuum, this is the algorithm that calculates the most efficient route to clean the entire floor without missing spots, while avoiding obstacles like couch legs and pet toys. For an autonomous vehicle, this