Sensors & actuators in robotics

A robot is, at its heart, a machine that senses its world, decides what to do, and acts on it. The components that handle the sensing and the acting are sensors and actuators, and they sit at opposite ends of the robot’s electronics. Everything else, the processing in between, exists to connect them.

Two opposite jobs

A sensor converts a physical quantity, light, distance, rotation, acceleration, temperature, into an electrical signal the controller can read. An actuator does the reverse: it converts electrical commands into physical action, most often motion. One brings information in; the other pushes effects out.

These two are tied together by the perceive-think-act loop, the fundamental cycle of any robot:

1. Perceive — read the sensors.
2. Think — process the data and decide what to do.
3. Act — drive the actuators.

Then it repeats, often many times per second. Because the loop runs continuously, the robot can react to a changing environment rather than blindly following a fixed script.

Sensing the world

Robots use a variety of sensors, each suited to different information:

• IMU (Inertial Measurement Unit). Combines a gyroscope (rotation rate) and an accelerometer (linear acceleration, including gravity) to track orientation and motion. Many add a magnetometer for heading. This is what keeps a drone or balancing robot upright.
• Rotary encoders. Count pulses as a shaft turns, reporting how far and how fast it has rotated. They give the precise position and speed feedback needed for accurate motor control and for odometry (estimating how far a robot has driven).
• Ultrasonic sensors. Emit a burst of ultrasound and time the returning echo. Using the speed of sound (about 343 m/s in air), distance follows from the round-trip time. Cheap and reliable for obstacle detection.
• Infrared (IR) sensors. Detect reflected infrared light for short-range proximity, obstacle detection and line following (telling dark from light surfaces). Inexpensive, but sensitive to ambient light and surface colour.
• Cameras. Provide rich 2D visual data. With computer vision a robot can recognise objects, read markers, follow lines and navigate; stereo cameras even estimate distance. The trade-off is the processing power vision requires.

A useful distinction is active versus passive sensors. An active sensor emits energy and measures what returns (ultrasonic, IR rangefinder, lidar). A passive sensor only receives existing energy (a plain camera, a microphone, a thermometer).

Acting on the world

Most actuators in robotics are motors, and the three common types serve different needs:

• DC motors. A brushed DC motor spins continuously when given voltage, with speed roughly proportional to that voltage. Simple and cheap, ideal for driving wheels, but it needs an encoder if you want precise position or speed.
• Servo motors. A hobby servo moves to and holds a commanded angle (often 0 to 180 degrees) using built-in position feedback, and is controlled by a PWM signal. Perfect for steering, robot-arm joints and grippers, where exact position matters.
• Stepper motors. A stepper moves in fixed, equal increments (steps). By counting steps it can be positioned precisely without a feedback sensor, which is why 3D printers and CNC machines rely on them for repeatable open-loop motion.

Closing the loop

The real power comes when sensors and actuators work together in a feedback loop. An actuator acts, a sensor measures the actual result, and the controller corrects any error. For example, a wheel motor is told to hold a target speed; its encoder reports the actual speed; if the wheel is slow (perhaps climbing a slope) the controller increases the drive. This continuous correction is what makes robots accurate and robust in the messy real world.

Open-loop control acts without checking the result; closed-loop control measures and corrects. Most precise robotics is closed-loop.

The bigger picture

Seen this way, a robot is a perceive-think-act loop wrapped around real hardware: a suite of sensors feeding data in, a controller deciding, and actuators carrying the decisions out, with feedback stitching the two ends together. Pick the right sensor for the information you need and the right actuator for the motion you want, connect them through a tight control loop, and you have the foundation of essentially every robotic system, from a line-following toy to an autonomous vehicle.