MODULE 04 · 2+ HOURS · ROBOTICS & AUTOMATION

Robotics &
Intelligent Automation

Plain-language tour: what robots are, how they see and move, and how factories use smart machines — with everyday examples.

See → think → move
Sensors & safety
10 Quiz Questions
8 Real-World Examples
10 Quiz Questions
🦾 Robot + AI map 📡 Sensors 🧭 Decisions 🏭 Automation
↓ scroll to begin

Machines that sense, think a little, and move

A robot is not magic. It is a machine that looks at the world, decides what to do, and moves — again and again, like a loop.

💡 Remember from earlier modules: AI can help a robot recognise things (a box, a person, a line on the floor). The robot still needs wheels, motors, and safety rules to act safely.

Simple definition: A robot is a machine that can sense the world, make choices, and take action — often without a human holding a remote control every second.

Everyday examples:

The robot loop (like a game):

  1. See — cameras, distance sensors, bumpers.
  2. Understand — “There is a wall 30 cm ahead.”
  3. Decide — “Turn left.”
  4. Move — motors spin, wheels turn.
  5. World changes — you moved, so you look again. Repeat!
THE ROBOT LOOP See Think Move Repeat The world changes after you move — so the loop never really stops.

Figure — Same idea as a thermostat: check temperature → adjust heater → check again.

Robots vs normal machines

MachineWhat it does
Washing machineRuns one fixed program you picked — not really “looking” at clothes.
Roomba-style vacuumChanges path when it sees a chair leg — reacts to the room.
Car with cruise controlKeeps speed — simple control, not full “robot.”
Self-driving car (goal)Sees lanes, signs, people — many decisions per second.
Story — delivery robot on campus: It uses a camera to see corridor posters and people. If someone blocks the path, it stops or goes around. AI might help it recognise “person” vs “bin,” but someone still chose how fast it may drive and where cameras sit.

Parts of a robot (what each piece does)

INSIDE A TYPICAL MOBILE ROBOT Computer / brain runs programs + AI Battery power for motors Camera Wheels + motors (move) Sensors in → brain decides → motors out

Figure — Sense, think, act: the same idea for almost every robot.

How much can it do alone?

Human drives everything Remote-control toy car — you are the brain.
Robot helps Car parks itself but you drive on the road.
Robot does the task Factory arm repeats the same pick-and-place all day.
Mostly alone (in a safe area) Warehouse robot follows floor lines; humans stay in marked zones.
Try it · Name three “robots” you have seen

For each one, say: Does it move? What does it sense? Who stops it if something goes wrong?


How robots “feel” the world

Sensors are the robot’s eyes, ears, and touch. Without good sensing, even the smartest AI is guessing in the dark.

Two big groups (easy names):

Combining sensors is like using both eyes and your inner-ear balance when you walk on ice — one sensor alone can fool you (mirrors confuse cameras; glass confuses some laser sensors).

TWO GROUPS OF SENSORS Outside — look at world camera · LiDAR · bumper Inside — feel own body wheel count · tilt · heat

Figure — Use more than one type when possible.

HOW ANY SENSOR WORKS World Sensor Computer Move / beep Sensor turns the world into numbers; software chooses the action.

Figure — Same chain for every sensor type.

Types of sensors — what they measure and where they are used

Sensor typeWhat it measuresUsed forExample
CameraLight and colour (picture)See people, boxes, defectsPhone, vacuum bot, CCTV
UltrasonicDistance (sound echo)How far to a wallHobby robot, parking beep
LiDARDistance (laser)360° room mapRobot vacuum
BumperPhysical touchStop when hitVacuum front, elevator
IMUTilt and spinStay balancedDrone, phone screen rotate
Wheel encoderWheel turnsHow far movedWheeled robots
GPSOutdoor positionMaps and routesDelivery drone
MicrophoneSoundVoice, alarmsSmart speaker
TemperatureHeatOverheat, cold trucksFactory motor, fridge sensor
Infrared proximityObject nearbyHands-free tap, dryerBathroom sensor
Example — shiny package in a factory: The camera kept missing defects on glossy plastic until workers added softer lights. The AI model did not change first — the seeing got better. Lesson: fix sensing before blaming “bad AI.”

Mapping while moving (SLAM — say it “slam”)

SLAM — DRAW MAP + FIND YOUR SPOT Where am I? What does the room look like? Vacuum bots do this in your home.

Figure — Mapping while moving (say “slam”).

When a robot enters a new building, it must answer two questions at once: Where am I? and What does this place look like? That is SLAM (simultaneous localisation and mapping). Think of walking through a dark hotel with a phone flashlight — you build a mental map as you go.

Think about it · What could go wrong?

For each, name a second sensor or a simple safety rule (slow down, stop, human button).


How robots choose what to do next

After sensing, the robot needs a plan: go forward, turn, stop, pick up. Some paths are drawn on a map; some are learned from practice.

Path planning (like GPS for robots): Given a map and a goal (“go to charging dock”), the robot finds a route that avoids tables and walls. Classic methods have names like A* — you only need the idea: search for a safe path.

Learning from trial and error: Some robots practise in simulation or a safe pen — try walking, get a score (“you fell”), try again. That is the spirit of reinforcement learning, but real robots still need safety limits so practice does not hurt anyone.

Follow the human: Many “smart” factory arms are first moved by a worker’s hand, then repeat that motion. Less fancy, very common.

Ways robots decide — types and when to use them

Way to decideHow it worksUsed forExample
Simple rulesIf sensor says X, do YVery predictable jobsBumper → stop
Map + pathSafe route on a mapWarehouse, hospital botGo to charger
Follow a lineCamera sees floor tapeSimple factory cartsYellow line follower
Teach by handHuman moves arm, robot copiesWelding, paintingFactory teach mode
Practice with scoresTry, get points, repeatResearch, gamesWalk in simulation
FIND A PATH AROUND A CHAIR chair start goal Curved path = planner avoided the obstacle.

Figure — Real robots also respect “do not enter” zones and speed limits.

Example — night patrol robot: Each night it must visit charging stations. Sensors: camera + wheel counters. Decision: use a map planner for the route; if battery is low, skip optional stops. You do not need a Hollywood AI brain — you need a clear goal and safe rules.
Quick compare · Two ways to decide
WayGood whenWatch out
Map + rulesBuilding layout stays similarNew furniture blocks paths
Learned behaviourMessy, changing tasks (research)Needs lots of safe practice; hard to explain mistakes

Factories and “smart” lines

Automation means machines do repetitive work. AI often helps with “is this part good or bad?” while old-fashioned safety systems still stop the line if something looks dangerous.

SIMPLE FACTORY LINE Parts in Camera + AIgood or bad? PLC rules Belt out E-stop

Figure — AI sees defects; PLC and red button keep people safe.

Factory parts — what each one does

PartJobUsed for
Robot armPick, place, weldSame motion all day
Conveyor beltMove parts alongLinking machines
Camera + AISpot scratches, wrong labelsVisual checks
PLC boxSafety rules, fast stopsButtons, fences, limits
CobotWork near people, slowerAssembly beside worker
Emergency stopCut power when pressedAlways required

PLC (think: factory brain in a rugged box): A small computer that reads buttons and sensors and turns motors on/off very reliably, many times per second. It follows fixed rules: if emergency button pressed, then stop everything.

Where AI fits: Cameras check scratches, dents, or wrong labels — things that are hard to write as simple rules. The AI suggests “bad part,” but a safety system still decides whether the machine may move.

Cobot = collaborative robot — designed to work near people, often slower and with force limits so a bump hurts less.

Safety rules (always on) stop button · fences · limits + AI vision (helps decide) good part? bad part? Only if BOTH agree → conveyor may run

Figure — AI does not replace the emergency stop button.

Example — packaging line: A camera spots a torn label. The AI flags it. The PLC stops the belt and flashes a light. A worker fixes the roll. Without the worker and the stop button, one wrong label could ship thousands of bad boxes.
Fairness reminder: If a vision system is trained only on one skin tone or one glove colour, it might pause the line unfairly. Good data and testing matter — you will see this again in Module 7.
Wrap-up · What should you remember?
  1. Robots loop: see → think → move → repeat.
  2. Sensors must match the job; fix lighting before buying a bigger AI model.
  3. Plans can be maps or learned — safety rules come first.
  4. Factories combine reliable stops with AI for messy visual checks.

Quick Knowledge Check

10 easy questions on robots, sensors, and factory helpers. Instant feedback on every answer.

Score: 0 / 0

Key Takeaways

Module 4 in short: robots sense, decide, and act — and people still matter for safety.

↑ Back to Top
📚 Further reading:
• ROS documentation — docs.ros.org
• IFR World Robotics reports — ifr.org
• Open Robotics / Gazebo simulation — gazebosim.org