Embedded System Design Cycle Questions and Answers

Answer: C

The controller runs its production firmware on its production hardware while a real time simulator pretends to be the engine, the motor or the vehicle. Faults that would destroy real equipment can then be injected repeatedly and safely.

Answer: D

Correct answer: dangerous and rare conditions to be tested repeatedly

Hardware in the loop runs the real controller against a simulated plant, so a fault condition that would destroy equipment or endanger someone can be injected as often as required, and a rare combination can be reproduced exactly. It also lets testing begin before the real plant exists, which is often what makes the schedule work.

Answer: A

Drivers, communication stacks and much of the application can be written and tested months before the first custom board arrives, which takes the software off the critical path. Its power and emissions bear no relation to the final product's.

Answer: B

Power first, because everything else can destroy the board if a rail is wrong; then the debug connection, which is the tool for everything after; then the simplest possible sign of life; and only then the peripherals one at a time.

Answer: C

A reversed regulator or a shorted rail can destroy every device on the board the moment power is applied, and it costs nothing to check with a meter first. Current-limiting the bench supply for the first power-up is the companion precaution.

Answer: D

Correct answer: a fault is then attributable to the one thing just changed

Bringing up one peripheral at a time keeps the number of possible causes at one. Enable six at once and a failure could come from any of them or from an interaction, and the search space is far larger than six. It is the same discipline as changing one variable at a time in an experiment, and it costs less time than it appears to.

Answer: A

Shipping an image nobody can reconstruct makes a field fault impossible to investigate, so the toolchain version, the flags and the library versions must be recorded with the source. Certification generally requires it outright.

Answer: B

Optimisation, undefined behaviour and library implementations all vary between versions, and code that works with one may fail with another. Pinning the version, and building in a container or a recorded environment, is what makes an investigation possible years later.

Answer: C

Correct answer: builds and tests every change, so a break is found at once

The value is in the immediacy: a defect found within minutes is attributable to a change the author still has in mind, while one found weeks later must be searched for. It also keeps the build itself working, which prevents the long integration phase that appears when everyone has been building only their own part. It does not replace review or on-target testing, both of which catch different things.

Answer: D

Compiling the algorithm for the host with the hardware behind a mocked interface gives fast feedback on the logic, which is where most bugs live. It says nothing about register access or timing, which on-target tests must cover.