Computer Science
Automated Systems, Robotics and AI
- 1.
An automated greenhouse turns irrigation on when soil moisture is below a stored threshold. Explain the roles of a sensor, microprocessor and actuator, then explain why the system must keep taking readings after the pump starts.
[4 marks] · no calculator - 2.
A warehouse proposes automated lighting controlled by movement sensors. Explain two benefits specific to this scenario, one disadvantage if sensor readings fail, and a practical mitigation that preserves safe access.
[4 marks] · no calculator - 3.
A programmable mobile warehouse robot has a frame, distance sensors, a microprocessor and motor-driven wheels. Explain why it is a robot by linking three stated features to robot characteristics. Explain one advantage over a worker entering a hazardous aisle and one limitation requiring a safe fallback.
[5 marks] · no calculator - 4.
A hypothetical greenhouse expert system stores plant facts and rules: IF leaves are wilted AND soil is dry THEN suggest watering; IF leaves are wilted AND soil is wet THEN suggest checking roots. A user enters 'wilted' and 'wet'. Identify the recommendation and explain the roles of the knowledge base, rule base, inference engine and user interface.
[5 marks] · no calculator - 5.
A hypothetical machine-learning system sorts pictures of recyclable objects after training on labelled pictures of glass and plastic. Explain how training differs from manually writing a rule for every picture. Describe how additional representative labelled examples could improve adaptation, and explain why testing on the training pictures alone is inadequate.
[4 marks] · no calculator - 6.
Compare two hypothetical route helpers. System A follows a stored set of expert rules using road-closure facts. System B adjusts predictions from new labelled travel-time records. Identify the expert-system and machine-learning approaches, explain how each uses data to produce an output, and explain why 'all AI must learn continuously' is incorrect.
[4 marks] · no calculator - 7.
A greenhouse controller records two sensor messages in a text file named Readings.txt. The file exists and contains exactly two lines, in order: 'moisture low' and 'moisture normal'. Write Cambridge-style pseudocode that opens this file for reading, reads both lines one at a time and writes each unchanged to a new file named Audit.txt, then closes both files. Declare a suitable variable and use a two-iteration loop. Explain why the controller keeps a file instead of retaining these messages only in RAM, and state what opening an existing Audit.txt FOR WRITE would do to its old contents. Do not control any real equipment.
[6 marks] · no calculator - 8.
An automated-robot logic tester applies the input pairs A,B in the order 00,01,10,11. Create a complete table with one row per pair and output columns A AND B, A OR B, A NAND B, A NOR B, A XOR B and NOT A. Give every output as 0 or 1; NOT A depends only on A even though the tester varies B. The tester is checking gate functions, not authorising a real robot to move.
[6 marks] · no calculator - 9.
A robot transmits seven data bits 1011001 with one even-parity bit appended on the right. Determine the parity bit and transmitted eight-bit word. A receiver gets 10110110; decide whether parity detects an error. A different corrupted word from the same transmission is 10010000; explain why parity accepts it despite the changed data. Describe how an ARQ protocol uses positive/negative acknowledgements and a timeout to request or trigger retransmission, and explain why ARQ does not make this parity check able to detect every corruption.
[6 marks] · no calculator - 10.
A simplified word-addressed Von Neumann processor in an automated controller has PC=120 and ACC=5. Memory address 120 contains the single-word instruction ADD 350, meaning add the value stored at address 350 to ACC. Memory address 350 contains 7. There is no branch, and each instruction occupies one word. Describe the fetch-decode-execute sequence, tracing PC, MAR, MDR, CIR and ACC where relevant. Explain how the address, data and control buses are used, and give the final PC and ACC. The symbolic instruction is supplied; no machine-code encoding is required.
[6 marks] · no calculator - 11.
A hypothetical fleet of warehouse robots pays for charging using a digital currency that exists only electronically. Its specified blockchain protocol stores time-stamped transaction records in blocks; each new block contains a hash link to the preceding block, and multiple participating nodes keep copies and validate proposed histories under shared rules. Explain what makes the payment a digital-currency transaction, how the ledger records transfers, and how hash links and shared validation make an attempted change to an older transfer evident and harder to accept. Explain why this does not guarantee that an original charge-meter reading or payment recipient was correct. No real transaction is requested.
[5 marks] · no calculator - 12.
Write one complete algorithm in Cambridge-style pseudocode or a high-level programming language for this hypothetical controller audit. It must not operate real equipment. Input an integer N, the number of stations, and re-input until 1<=N<=20. For each station in input order, input a nonempty name (re-input an empty name) and exactly three numeric temperature readings; re-input each reading until it lies in 0..100 inclusive. Input is already of the stated integer/numeric type, so text parsing is not required. Store names in a one-dimensional array and all valid temperatures in a two-dimensional array. Calculate and store each station's mean in a one-dimensional array. A station has status ALERT if at least two of its three readings are >=70 OR any one reading is >=90; otherwise its status is NORMAL. Output each station's name, mean and labelled status. Count stations with ALERT once per station, not once per reading. Finally output that count and the name and mean of the station with the highest mean; retain the first input station if means tie. Declare suitable data structures and initialise totals and flags. Arrays can be sized for 20 stations. This original 15-mark practice scenario has an indicative analytic scheme, not an official board mark award.
[15 marks] · no calculator