Optimization of test data management using unsupervised learning

Challenge

In the context of vehicle tests for quality assurance, a manufacturer selects a variety of vehicles. Since the number of possible configurations is very large in this case, the aim is to determine a subset of specific vehicle configurations that is as representative as possible. This makes it possible to achieve the highest possible test coverage with regard to specific equipment features. The previous manual process is time-consuming and does not achieve optimal test coverage.

Solution

Using Unsupervised Learning (clustering) similar vehicle configurations are recognized in the test pool and combined into clusters. One representative configuration is identified per group.

Benefit

• A proposal function of the AI makes a systematic determination of the subset of vehicles possible
• The test coverage is maximized for a given number of vehicle configurations (cluster)