On a complex piece of automotive software, testing newly deployed code is vital. With so many interdependencies between each function, those tests ensure any changes don’t negatively affect other areas of the system. However, with the amount of code and its complexity increasing with every build, running every single test for every version of the software can take hours.

While this is a necessity for the system software before it is released for production or before OTA updates are to deployed, running all the tests for daily builds while still in development is hugely inefficient — especially if only a specific function has been updated.

Identify all affected systems

Even when a new version of the software seems to only update one specific function, it’s important to identify which other systems could be affected. The Line-of-Code IntelligenceTM technology developed by Aurora Labs maps all interdependencies to better understand the effect of any changes on other functions within the vehicle.

 

Survey
The 2023 4th Annual Automotive Software Strategy Survey
Check out the latest insights from the industry from our 4th Annual Automotive Software survey conducted together with TechInsights. Don’t miss the latest trends on eMobility, OTA updates and how AI can enhance the security and customer experience of EV and autonomous vehicles.
Read more

Eliminating test execution redundancy

To identify these dependencies, Aurora Labs’ Auto Detect adds a layer of artificial intelligence (AI) to the software testing process. This helps to detect exactly which tests are needed for each software function, enabling the tester to focus on and only run the required tests, eliminating the need to re-run all tests every time a new version is committed to the software repository.

For example, if a system has 100 tests, Auto Detect will be able to select the right ones based on exactly what’s been changed by an update — including which other functions might be affected by the new code. This means that of those 100 tests, only 20 might need to run. This saves valuable time by ensuring only the necessary tests are performed.

By adding this AI layer to your development process, it’s possible to decrease the time it takes to run tests. On top of the increased efficiency, you improve quality by ensuring dependent functions are also tested alongside those that have been directly updated — without the need to run redundant tests.

If you’d like to find out more about how Auto Detect can improve the automotive software testing process, find out more here.