Feature-Driven Rapid Prototyping of Test-Sequences for Sensor Characterization in the Laboratory

Nowadays, testing in the development of consumer sensors such as low-cost inertial measurment units (IMUs) requires not only highly flexible laboratory infrastructure, but also software artifacts that implement the intended test sequences on increasingly complex setups. To overcome this significant challenge across multiple domains, we propose using feature models to describe the contextual variability under which a test can be run. Consequently, a configuration of such a model is used to specify a single test and compute a sequence of context enabling actions, that form the core of a generated extensible code structure for the required test sequences, hereby allowing sensor experts to rapidly specify and generate implementations of test sequences for complex setups. We evaluate our approach on the development of noise-measurement sequences for a consumer IMU and show how feature models can be used as test specification and for subsequent code generation.