YOUR REVISION ROUTE
Read a summary, close your notes and explain the idea. Use the cards to select weak points and build a revision list linked to the lessons.
6.1 | AUTOMATED SYSTEMS
Automated systems perform programmed operations with reduced manual involvement. Sensors provide information, processors execute rules and outputs affect the task or report results.
6.1.1 | MEASURE, PROCESS, ACT
| Component | Role | Greenhouse example |
|---|---|---|
| Sensor | Measure a physical quantity | Temperature reading |
| Microprocessor | Execute instructions and determine outputs | Compare reading with programmed limits |
| Actuator | Produce a physical action | Motor drives the fan |
Feedback: new readings after an action guide later decisions. The controller repeats the input–process–output cycle rather than assuming one reading stays valid forever.
An analogue signal may need an ADC to create a digital representation. Some sensors already provide digital information, so a separate external ADC is not universal.
Worked rule: fan on above 28°C, off below 26°C, keep state between. Starting off, readings 25, 29, 27, 25 give off, on, on, off. These are illustrative thresholds. The band is hysteresis, reducing rapid switching near one limit.
6.1.2 | ALL SEVEN AUTOMATION SCENARIOS
| Area | Example | Benefit and limitation |
|---|---|---|
| Industry | Product checking and sorting | Consistent throughput; faulty rules can repeat errors |
| Transport | Responsive traffic signals | Adjust to demand; detection faults and safety constraints matter |
| Agriculture | Irrigation using soil readings | Target watering; readings may not represent every area |
| Weather | Regular instrument logging | Remote or continuous data; calibration and outages affect it |
| Gaming | Motion-responsive gameplay | Engaging input; tracking error or delay can frustrate |
| Lighting | Occupancy and brightness control | Potential energy savings; stationary users may be missed |
| Science | Controlled experiment and logging | Repeatability; systematic measurement errors can remain |
Monitoring collects and reports information. Control uses readings to affect a process. A weather station does not automatically control the weather.
Precise repeated readings can still be inaccurate if biased. Automated does not mean maintenance-free, internet-connected or AI-based.
6.2 | ROBOTICS
Robotics is a branch of computer science incorporating the design, construction and operation of robots. Factory equipment, domestic robots and drones are syllabus examples.
6.2.1–2 | ROBOTICS AND CHARACTERISTICS
| Idea | Remember |
|---|---|
| Design | Choose structure, sensing, actuation and control to meet requirements |
| Construction | Build and integrate the system |
| Operation | Run, control, test and maintain it |
| Mechanical structure | Support the physical components and movement |
| Electrical components | Sensors, processors, actuators and suitable power |
| Programmable | Instructions specify sequences and responses |
A fixed arm can be a robot; it does not need wheels, legs or a human shape. An operator can supply commands while programmed control functions run.
A program cannot remove every physical constraint. Reach, payload, movement range and suitable power still matter.
6.2.3 | ALL SIX ROBOT ROLE AREAS
| Area | Role | Contextual trade-off |
|---|---|---|
| Industry | Handle parts or tools | Repeatability and exposure reduction; integration cost and downtime |
| Transport | Move items | Less repeated carrying; route, load and battery limits |
| Agriculture | Targeted physical crop work | Labour saving; variable conditions and identification errors |
| Medicine | Assist instruments or rehabilitation | Controlled physical assistance; professional oversight and training |
| Domestic | Vacuum or mow | Convenience; layout limits and upkeep |
| Entertainment | Perform or interact | Engaging physical behaviour; maintenance and audience safety |
Evaluate: identify the requirement, explain a relevant benefit, explain a limitation and conclude for that situation.
For predictable warehouse routes, robots may reduce manual carrying. However, blocked paths and charging can disrupt deliveries, so suitability depends on route planning and support.
6.3 | ARTIFICIAL INTELLIGENCE
AI is a branch of computer science dealing with the simulation of intelligent behaviours by computers. Simulation does not establish human consciousness or emotion.
6.3.1–2 | DEFINITION AND CHARACTERISTICS
The specification describes data collection and rules for using that data, the ability to reason, and the fact that AI can include learning and adaptation.
| Characteristic | Meaning |
|---|---|
| Data | Relevant information used by the system |
| Rules / method | Relationships determining how information produces results |
| Reasoning | Use information and relationships to reach a conclusion |
| Learning | Adapt a model, processes or data from training information |
| Adaptation | A change that affects later behaviour |
A fixed expert system can infer conclusions without learning. A trained model can give different results for different inputs without changing itself.
Automation, robotics and AI overlap but are not identical: a timer can automate, a robot can follow fixed instructions, and AI can run without a physical robot.
6.3.3 | EXPERT-SYSTEM COMPONENTS
| Component | Role |
|---|---|
| Knowledge base | Store domain facts and expert knowledge |
| Rule base | Store conditions and conclusions |
| Inference engine | Apply knowledge and rules to available information |
| Interface | Gather user inputs and present questions/results |
Worked rule: IF power is present AND paper is absent THEN recommend checking the printer’s paper tray. The interface obtains answers; the engine checks them against the rule; the interface presents the result.
The recommendation depends on the facts and rules. It does not prove this is the only possible fault. A wrong answer or missing knowledge can lead to an unsuitable conclusion.
6.3.3 | MACHINE LEARNING OPERATION
Machine learning allows a program to automatically adapt its processes and/or data. In an image-classification example, training adjusts a model using suitable examples and labels.
| Stage | Role |
|---|---|
| Collect and prepare | Obtain representative examples and check their quality |
| Train | Adapt the model through a chosen process |
| Evaluate | Check performance on suitable examples outside fitting |
| Inference | Use the trained model on a new input |
| Review | Investigate mistakes and decide on changes |
A model that fits familiar images may struggle with new backgrounds or poorly represented objects. More data does not guarantee improvement if labels or representation are unsuitable.
Expert systems use explicit knowledge and rules; machine-learning models are adapted from data. Real systems can combine both.
COMMON MISTAKES
- Calling the sensor the actuator or saying the processor creates the physical force.
- Assuming automated weather monitoring controls the weather.
- Claiming every robot is humanoid, autonomous or AI-based.
- Listing benefits without explaining a scenario consequence.
- Confusing knowledge storage with the inference engine’s reasoning.
- Assuming every AI learns continuously.
- Treating a new output as proof that the model changed.
- Using training success as proof of correct future results.
READY TO TEST YOURSELF?
Explain a control loop, evaluate a robot role, define AI, identify expert-system components and distinguish training from inference. Select cards where your explanation is still uncertain.