Control systems and translation (HL) – IB Computer Science exam questions with mark scheme

1CoreHL only3 marks

A home security system has a passive infrared (PIR) motion detector in the hallway. When the system is armed and movement is detected, a siren sounds.

(a) Identify the sensor and the actuator in this system. [2]

(b) State what is meant by a transducer. [1]

(a) Sensor: the PIR motion detector (accept: infrared sensor / motion sensor; also accept a door/window contact sensor if named as an additional sensor)(1)
(a) Actuator: the siren (accept: speaker / sounder / buzzer; also accept an electronic door lock or flashing light if described as part of the system; do not accept 'alarm' alone unless it is clear it means the sounding device)(1)
(b) A device that converts energy/a signal from one form into another (e.g. infrared radiation into an electrical signal, or an electrical signal into sound) (allow: converts a physical quantity into an electrical signal or vice versa; ignore examples without the idea of conversion)(1)
Answer: (a) Sensor: PIR motion detector; actuator: siren. (b) A device that converts one form of energy/signal into another.
Examiner tip: An actuator is the device that does something physical in response to the controller (siren, motor, valve), not the device that measures.

2CoreHL only4 marks

A washing machine controls each wash cycle using a microcontroller.

(a) In one model, the rinse stage simply runs the drum motor for 12 minutes and then stops. Outline why this stage is an example of open-loop control. [2]

(b) While filling, the machine keeps the water inlet valve open until a pressure sensor shows that the drum contains the required water level. Outline why this stage is an example of closed-loop control. [2]

(a) The output (motor running) is decided only by the input/setting (the timer of 12 minutes)(1)
(a) There is no feedback / the result (e.g. how well the clothes are rinsed, or motor speed) is not measured and does not affect what the controller does (accept: no sensor is used to monitor the output)(1)
(b) The pressure sensor measures the output/result of the process (the water level in the drum) and feeds it back to the controller (accept: feedback loop)(1)
(b) The controller compares the measured level with the required level/set point and uses this to decide whether to keep the valve open or close it (accept: the valve closes only when the target level is reached)(1)
Answer: (a) Runs for a fixed time with no feedback of the result. (b) Water level is measured and fed back; compared with the target to decide when to close the valve.
Examiner tip: The word 'feedback' alone is not enough: say what is measured and that it is compared with a set point to change the output.

3StandardHL only5 marks

The diagram shows the lane-keeping system of an autonomous vehicle. A camera measures the vehicle's position in the lane, a steering motor turns the wheels and a control algorithm runs on an onboard computer.

(a) Identify the components labelled A, B and C. [3]

(b) Describe the role of the feedback path in keeping the vehicle in the centre of the lane. [2]

A: the controller / onboard computer running the control algorithm (accept: processor / microcontroller)(1)
B: the actuator / steering motor(1)
C: the sensor / camera (accept: transducer)(1)
(b) The camera continuously measures the actual position of the vehicle and returns it to the comparator/controller, which compares it with the target position to give an error (accept: deviation / difference)(1)
(b) The controller uses the error to adjust the steering (e.g. steer left if the vehicle has drifted right) so the error is reduced towards zero; the process repeats continuously (accept: corrects for disturbances such as wind, road camber)(1)
Answer: A = controller (control algorithm), B = actuator (steering motor), C = sensor (camera). The measured position is fed back and compared with the target; the error is used to correct the steering continuously.
Examiner tip: Follow the arrows: whatever sits on the feedback path is the sensor, and whatever drives the process directly is the actuator.

Stuck on Control systems and translation (HL)?

Get it sorted in a free 30-minute session

We work through your weakest questions on this topic on Zoom. I reply the same day.

Free 30-min session on this topic Send me 5 questions a day

4StandardHL only5 marks

A thermostat uses the following control algorithm, run once every minute. The variable temp holds the latest sensor reading in °C.

setpoint = 20
heater_on = False

# repeated every minute
temp = read_sensor()
if temp < setpoint - 1:
    heater_on = True
elif temp > setpoint + 1:
    heater_on = False
set_heater(heater_on)

(a) The sensor returns these readings, in order: 18.5, 19.4, 20.6, 21.2, 20.8, 19.1, 18.9. The heater starts off. Trace the algorithm to give the state of the heater (ON or OFF) after each reading. [3]

(b) Explain why the algorithm uses a band of 1 °C either side of the set point instead of simply switching the heater on whenever temp < setpoint and off otherwise. [2]

(a) 18.5 → ON and 19.4 → ON (heater stays on because 19.4 is inside the band)(1)
(a) 20.6 → ON and 21.2 → OFF(1)
(a) 20.8 → OFF, 19.1 → OFF and 18.9 → ON (follow through one earlier error in the sequence)(1)
(b) With a single switching point the temperature hovers around 20 °C, so small fluctuations/sensor noise would make the heater switch on and off very frequently (accept: rapid cycling / chattering)(1)
(b) The band (hysteresis) means the heater changes state less often, reducing wear on the heater/relay and energy wasted on repeated start-ups (accept: more stable control / longer component life)(1)
Answer: (a) ON, ON, ON, OFF, OFF, OFF, ON. (b) Prevents rapid on/off switching near 20 °C caused by small fluctuations, reducing wear and energy waste.
Examiner tip: When a reading falls inside the band neither condition is true, so the heater keeps its previous state: it does not switch off.

5StandardHL only4 marks

A lift in a four-storey office building is run by a control system.

Explain how sensors and feedback are used to ensure that:

(a) the lift stops level with the floor that a passenger has requested; [2]

(b) the doors do not close on a passenger who is standing in the doorway. [2]

(a) Position sensors (e.g. floor/limit switches, magnetic or optical sensors, or a shaft encoder) detect where the lift car is and send this to the controller(1)
(a) The controller compares the car's position with the requested floor and slows then stops the hoist motor when they match / when the floor-level sensor is triggered (accept: motor speed reduced as the target floor approaches)(1)
(b) An infrared light curtain / beam sensor (accept: pressure sensor on the door edge) detects an obstruction in the doorway while the doors are closing(1)
(b) The controller receives this signal and immediately stops/reverses the door motor so the doors reopen; the doors only close once the sensor shows the doorway is clear(1)
Answer: (a) Position sensors report the car's floor; controller compares with requested floor and stops the motor when level. (b) Light curtain/pressure sensor detects a person; controller stops/reverses the door motor.
Examiner tip: Name the sensor and then say what the controller does with its reading; marks are for the complete sensor → controller → actuator chain.

6HardHL only6 marks

A greenhouse irrigation system measures soil moisture continuously. The controller switches a water pump on when the moisture reading falls to 30% or below, and switches it off when the reading reaches 45% or above. The graph shows the sensor readings over 16 hours.

(a) Using the graph, state:

(i) the two times at which the pump was switched on; [1]

(ii) the time at which the pump was switched off. [1]

(b) The moisture reading continued to fall for about an hour after the pump was first switched on. Explain why. [2]

(a)(i) 4 hours and 14 hours (both needed)(1)
(a)(ii) 7 hours(1)
(b) It takes time for the water to soak through the soil to reach the sensor (accept: delay/lag between the actuator acting and the sensor detecting the change)(1)
(b) Meanwhile the soil continues to lose water (through plants taking it up / evaporation), so the reading carries on falling until the water arrives (ignore: simply restating that the reading fell from 30% to 28%)(1)
(c) The output of the process (soil moisture) is measured by the sensor and fed back to the controller(1)
(c) The controller compares the reading with the thresholds/set points to decide whether the pump is on or off, so the pump's operation depends on the result it produced (reject: answers that only say 'it uses a sensor')(1)
Answer: (a)(i) 4 h and 14 h; (ii) 7 h. (b) Water takes time to reach the sensor while the soil keeps drying (lag). (c) Moisture is measured and fed back and compared with thresholds to control the pump.
Examiner tip: Read the switching times where the line meets the threshold lines, not at the lowest or highest points of the curve.

7CoreHL only3 marks

A programmer writes an application in Java. The compiler translates it into bytecode, which is distributed to users running Windows, macOS and Linux.

(a) State what is meant by bytecode. [1]

(b) Outline why distributing the program as bytecode makes it portable. [2]

(a) An intermediate (low-level) instruction set produced by compiling source code that is not the machine code of any particular real processor; it is executed by a virtual machine (accept: platform-independent intermediate code)(1)
(b) The same bytecode file can be run on any platform/processor that has a virtual machine (e.g. JVM) installed (accept: write once, run anywhere)(1)
(b) The virtual machine on each platform interprets/JIT-compiles the bytecode into that machine's own native instructions, so the program does not need to be recompiled for each operating system/processor(1)
Answer: (a) Intermediate code executed by a virtual machine, not native machine code. (b) Any platform with a VM can run the same file; the VM translates it into native code locally.
Examiner tip: Portability comes from the virtual machine on each platform; saying 'bytecode works on all computers' without mentioning the VM loses the mark.

8StandardHL only5 marks

A small studio is building a mobile game.

(a) While designing new levels, the developers change the level-scripting code many times an hour and test each change immediately. Explain why an interpreter is suitable for this work. [2]

(b) The physics engine must perform millions of calculations per second on a phone processor. Explain why a compiler is more suitable for this part of the game. [2]

(a) The code can be run straight away / without waiting for the whole program to be compiled after each change, so the edit–test cycle is fast(1)
(a) Execution stops at the line where an error occurs (often with the program state available), so errors are easy to locate and fix during testing (accept: line-by-line execution supports debugging)(1)
(b) The source code is translated into native machine code once, before it runs, so no translation work is done during execution / at run time(1)
(b) The compiled code therefore runs faster (and the compiler can optimise it for the target processor), which is needed for performance-critical calculations (accept: an interpreter would retranslate statements every time they execute, slowing the engine)(1)
(c) A compiler checks the whole program and reports all the (syntax) errors it finds before any of it runs / no executable is produced until they are fixed, whereas an interpreter reports an error only when it reaches the faulty statement, after earlier statements have already run (accept either half if the contrast is clear)(1)
Answer: (a) No compile wait, so changes can be tested instantly; stops at the error line for easy debugging. (b) Translated once to optimised native code so it runs faster with no run-time translation. (c) Compiler lists all errors before running; interpreter stops at the first error reached during execution.
Examiner tip: Tie each advantage to the scenario: 'faster' alone gets nothing; say what is faster (testing cycle vs execution) and why.

9StretchHL only6 marks

A bytecode virtual machine can either interpret every instruction or use just-in-time (JIT) compilation. For a particular loop, measurements show:

Mode	Behaviour
Interpreter only	Each iteration takes 50 µs.
With JIT	The first 1000 iterations are interpreted (50 µs each) while the VM detects that the loop is ‘hot’. It then spends 20 ms compiling the loop to native code. Every later iteration takes 2 µs.

(a) Calculate the total time taken for 10 000 iterations in each mode. Give your answers in ms. [2]

(b) Calculate the smallest number of iterations for which the JIT mode is faster overall. [2]

(c) Explain why the VM interprets the loop first instead of compiling all of the program to native code before it starts running. [2]

(a) Interpreter: 10 000 × 50 µs = 500 000 µs = 500 ms(1)
(a) JIT: 1000 × 50 + 20 000 + 9000 × 2 = 88 000 µs = 88 ms (allow 1 mark only if both values correct but left in µs)(1)
(b) Correct inequality/equation for n > 1000, e.g. 50n > 50 000 + 20 000 + 2(n − 1000) giving 48n > 68 000 (or n > 1416.7)(1)
(b) 1417 iterations (do not accept 1416 or 1416.7)(1)
(c) Compiling the whole program before running would cause a long start-up delay / the user waits before anything happens(1)
(c) Much of the code runs only once or rarely, so compiling it would cost more time than it saves; the VM compiles only frequently executed (hot) code, where the one-off compilation cost is repaid by faster later iterations (accept: profiling while interpreting lets the JIT optimise for how the code is actually used)(1)
Answer: (a) Interpreter 500 ms; JIT 88 ms. (b) 1417 iterations. (c) Avoids a long start-up delay and wastes no compile time on code that runs rarely; only hot code repays the compile cost.
Examiner tip: Remember the first 1000 iterations cost the same in both modes, so only iterations after 1000 benefit from the 48 µs saving.

10StretchHL only6 marks

A company is developing a photo-editing application that must run on Windows laptops, Apple computers and Linux workstations. It is considering two approaches:

Approach 1: compile the source code separately into native machine code for each platform.
Approach 2: compile the source code once into bytecode, which users run on a virtual machine that uses JIT compilation.

Evaluate the two approaches for this application. [6]

Approach 1: native code runs at maximum speed with no run-time translation; useful for heavy image-processing filters on large photos(1)
Approach 1: compiler can optimise for each target processor (e.g. use processor-specific instruction sets such as SIMD/vector extensions); no virtual machine needs to be installed by users(1)
Approach 1 limitation: a separate build (and possibly code changes) is needed for each platform/processor, so development, testing and updates take longer and cost more(1)
Approach 1: source code is not distributed and machine code is harder to reverse-engineer than bytecode (protection of intellectual property)(1)
Approach 2: one bytecode file runs on every platform that has the VM, so one build to test, distribute and update (portability)(1)
Approach 2: JIT compiles frequently used (hot) code such as filter loops to native code at run time, so performance can approach native speed after warm-up(1)
Approach 2 limitation: start-up/warm-up delay, extra memory used by the VM and JIT, and users must have a suitable VM installed(1)
Approach 2: the VM can provide extra run-time checks (e.g. memory safety/sandboxing) that help security and error detection(1)
A justified conclusion linked to the scenario, e.g. Approach 2 is preferable because supporting three platforms cheaply matters more than a small loss of speed, or Approach 1 if peak performance on very large images is the priority(1)

Level	Marks	What it looks like
Level 3	5–6	Strengths and limitations of both approaches are explained and clearly linked to a cross-platform photo editor; reasoning is accurate and leads to a justified conclusion.
Level 2	3–4	Relevant points about both approaches with some explanation, but the evaluation is uneven, only partly linked to the scenario, or the conclusion is weak or unsupported.
Level 1	1–2	Isolated, mainly descriptive points (e.g. 'compiled code is faster'), possibly about only one approach, with little or no link to the scenario and no conclusion.
No marks	0	Nothing relevant or worthy of credit.

Answer: Balanced appraisal: native compilation gives the best performance and no VM dependency but needs a separate build per platform; bytecode + JIT gives one portable build with near-native speed for hot code but has start-up overhead and needs a VM. Justified conclusion for a photo editor.
Examiner tip: 'Evaluate' needs strengths and limitations of BOTH approaches and a conclusion tied to the photo-editing scenario; a list of facts with no judgement stays in the middle band.

Stuck on this topic?

Book a free 30-minute session on WhatsApp and we'll find exactly which step is costing you the marks. Or build a timed paper on it.

Free 30-minute session